{"id": "video_109__q0000", "video_id": "video_109", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_109_t5s.mp4"], "extra_info": {"question": "How many athletes participating in the race were shown in the video while the host was introducing Noah Lyles before the race?"}, "reward_model_ground_truth": {"multi_choice": ["A.2", "B.3", "C.4", "D.5"], "correct_option": "D"}, "answer_text": "5", "reasoning_steps": [{"modality": "audio", "evidence": "host was introducing Noah Lyles in around 0:03", "inference": "the athletes show up in around 0:03 are required in the question."}, {"modality": "vision", "evidence": "Noah Lyles is in the left side of the frame", "inference": "he is the first one appeared"}, {"modality": "vision", "evidence": "an athletes in white and purple is in the frame", "inference": "he is the second one appeared"}, {"modality": "vision", "evidence": "an athletes with golden hair is in the frame", "inference": "he is the third one appeared"}, {"modality": "vision", "evidence": "an athletes in red is at the 4th lane", "inference": "he is the 4th one appeared"}, {"modality": "vision", "evidence": "an athletes in white is at the 3rd lane", "inference": "the answer is 5"}]}
{"id": "video_120__q0000", "video_id": "video_120", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_120_t5s.mp4"], "extra_info": {"question": "When Rachel complained about spending over $1,000, what color of clothes was the person next to the one who suggested returning the item wearing? "}, "reward_model_ground_truth": {"multi_choice": ["A.PURPLE", "B.RED", "C.BLUE", "D.GREY"], "correct_option": "B"}, "answer_text": "RED", "reasoning_steps": [{"modality": "audio", "evidence": "Rachel says'I don't know what I'm gonna do with this thing'at 2:11.", "inference": "She complains spending over $1,000 on this cat."}, {"modality": "audio", "evidence": "A woman says 'take it back to where you got it.' at 2:19.", "inference": "She suggests returning the item."}, {"modality": "vision", "evidence": "The woman suggests returning the item wearing a purple clothes.", "inference": "Identify the person who suggested returning the item."}, {"modality": "vision", "evidence": "the person next to this woman is a man wearing red shirt.", "inference": "The color is red."}]}
{"id": "video_137__q0000", "video_id": "video_137", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_137_t5s.mp4"], "extra_info": {"question": "How many singers sing incorrectly?"}, "reward_model_ground_truth": {"multi_choice": ["A. 4", "B. 6", "C. 5", "D. 3"], "correct_option": "C"}, "answer_text": "5", "reasoning_steps": [{"modality": "audio", "evidence": "The first on-screen person speaks before any singing begins.", "inference": "Note that subsequent entries are solo vocals presented one-by-one."}, {"modality": "vision", "evidence": "If singers are sing wrong lyrics, the subtitles at the bottom of the screeen will be circled by red color.", "inference": "We can judge whether the singers is sing incorrectly through red circle."}, {"modality": "vision", "evidence": "When singer #2 stands out, during his performance, some of the subtitles at the bottom of the screeen are circled by red color.", "inference": "Singer  #2 sings incorrectly."}, {"modality": "vision", "evidence": "When singer #3 stands out, during his performance, some of the subtitles at the bottom of the screeen are circled by red color.", "inference": "Singer  #3 sings incorrectly."}, {"modality": "vision", "evidence": "When singer #4 stands out, during his performance, some of the subtitles at the bottom of the screeen are circled by red color.", "inference": "Singer  #4 sings incorrectly."}, {"modality": "vision", "evidence": "When singer #5 stands out, during his performance, some of the subtitles at the bottom of the screeen are circled by red color.", "inference": "Singer  #5 sings incorrectly."}, {"modality": "vision", "evidence": "When singer #6 stands out, during his performance, some of the subtitles at the bottom of the screeen are circled by red color.", "inference": "Singer  #6 sings incorrectly.Since SIngers #2-6 sing incorrectly,the answer is five."}]}
{"id": "video_148__q0000", "video_id": "video_148", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_148_t5s.mp4"], "extra_info": {"question": "How many people (not including herself) have the same hair color as the person referred to as the 'fairy godmother'?"}, "reward_model_ground_truth": {"multi_choice": ["A 3", "B 4", "C 5", "D 6"], "correct_option": "A"}, "answer_text": "3", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:12, the woman was called 'fairy godmother'.", "inference": "Identify the person mentioned in the question."}, {"modality": "vision", "evidence": "At 0:11, the woman has black hair.", "inference": "We need to notice the people with black hair."}, {"modality": "vision", "evidence": "At 0:03, the girl has black hair.", "inference": "This is the first one."}, {"modality": "vision", "evidence": "At 1:39, the judge on the left has black hair.", "inference": "This is the second one."}, {"modality": "vision", "evidence": "At 1:39, the judge on the right has black hair.", "inference": "This is the third one."}]}
{"id": "video_150__q0000", "video_id": "video_150", "qid": 0, "question_type": "ego reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_150_t5s.mp4"], "extra_info": {"question": "What happened to me when the sirens faded away?"}, "reward_model_ground_truth": {"multi_choice": ["A. I threw the security guard away.", "B. A security guard appeared and chased me.", "C. I ran into a security guard.", "D. I was chased by two security guards."], "correct_option": "B"}, "answer_text": "A security guard appeared and chased me.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:01, the siren begins to sound.", "inference": "Remind people to notice the sirens."}, {"modality": "audio", "evidence": "At 1:00, the sirens faded away.", "inference": "Mark the point mentioned in the question that needs attention."}, {"modality": "vision", "evidence": "At 1:08, a security guard appears.", "inference": "This is the first part of the answer."}, {"modality": "vision", "evidence": "At 1:12, a security guard begins to chase me.", "inference": "This is the second part of the answer."}]}
{"id": "video_152__q0000", "video_id": "video_152", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_152_t5s.mp4"], "extra_info": {"question": "How many prison guards did I see before I shouted 'go' more than a dozen times?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2.", "B. 3.", "C. 4.", "D. 5."], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:16, 'I' see two prison guards.", "inference": "There are two prison guards."}, {"modality": "vision", "evidence": "At 2:17, 'I' see another two prison guards.", "inference": "There are four prison guards in total."}, {"modality": "audio", "evidence": "At 2:19, 'I' shout 'go' more a dozen time.", "inference": "This is the time point mentioned in the question."}]}
{"id": "video_153__q0000", "video_id": "video_153", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_153_t5s.mp4"], "extra_info": {"question": "Did I succeed in my challenge? Where did I run to?"}, "reward_model_ground_truth": {"multi_choice": ["A. Yes,Subway Exit.", "B. Yes,Inside the subway.", "C. No,Subway Exit.", "D. No,Inside the subway."], "correct_option": "B"}, "answer_text": "Yes,Inside the subway.", "reasoning_steps": [{"modality": "vision", "evidence": "I begin running at 1:06", "inference": "I am running to somewhere."}, {"modality": "vision", "evidence": "I stop running after get into the subway at 4:07.", "inference": "I ran inside the subway."}, {"modality": "audio", "evidence": "My friend says 'you made it'.", "inference": "I succeed in my challenge."}]}
{"id": "video_157__q0000", "video_id": "video_157", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_157_t5s.mp4"], "extra_info": {"question": "What does the sound effect that appears after I hear thank you mean?"}, "reward_model_ground_truth": {"multi_choice": ["A. Means he is grateful for my help.", "B. Means I may be stolen again.", "C. Means I'm happy to help others in need.", "D. Means he is happy to receive the money I sent back."], "correct_option": "B"}, "answer_text": "Means I may be stolen again.", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:07, I hear 'thank you'.", "inference": "This is the moment mentioned in the question."}, {"modality": "vision", "evidence": "At 4:07, I returned the recovered money to the homeless person.", "inference": "This is a good deed that I did."}, {"modality": "audio", "evidence": "At 4:09, a sound effect suddenly appears.", "inference": "Something must have happened."}, {"modality": "vision", "evidence": "At 4:10, the man who stole money from the homeless person has come back again.", "inference": "It indicates that I may be doomed."}]}
{"id": "video_175__q0000", "video_id": "video_175", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_175_t5s.mp4"], "extra_info": {"question": "Why did I throw down that toilet?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because the toilet is broken.", "B. To demonstrate the height here.", "C. To measure the height here.", "D. To evade the pursuing troops."], "correct_option": "B"}, "answer_text": "To demonstrate the height here.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:01, 'I' hold a toilet.", "inference": "Introduce the scene of the video."}, {"modality": "vision", "evidence": "At 0:04, 'I' throw down the toilet.", "inference": "Mark the point that needs attention about the question."}, {"modality": "audio", "evidence": "At 0:04, 'I' say 'I'm throwing it from such height. Well, I'll be down there in 20 seconds.'", "inference": "Prove the height that 'I' am going to challenge."}]}
{"id": "video_179__q0000", "video_id": "video_179", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_179_t5s.mp4"], "extra_info": {"question": "How many minutes were the shrimp dumplings heated in the microwave?"}, "reward_model_ground_truth": {"multi_choice": ["A. 3.", "B. 4.", "C. 5.", "D. 2."], "correct_option": "C"}, "answer_text": "5.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:58, the protagonist put a cage of shrimp dumplings into the microwave oven.", "inference": "He might mention the time to use the microwave around this point."}, {"modality": "audio", "evidence": "At 0:57, he said that'I'll microw it for 3 minutes.'.", "inference": "It indicates that the time he used the microwave to heat the shrimp dumplings is 3 minutes."}, {"modality": "audio", "evidence": "He said that 'Two more minutes.'.", "inference": "It indicates that he added another 2 minutes to heat the shrimp dumplings. So, the shrimp dumplings were heated in the microwave for a total of 5 minutes."}]}
{"id": "video_197__q0000", "video_id": "video_197", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_197_t5s.mp4"], "extra_info": {"question": "What is the animal's tone like when they meet in the tunnel?"}, "reward_model_ground_truth": {"multi_choice": ["A. a flustered tone", "B. a disgusted tone", "C. a polite tone.", "D. a rhetorical tone"], "correct_option": "C"}, "answer_text": "a polite tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 3:58, they meet in the tunnel.", "inference": "This is the time mentioned in the question."}, {"modality": "audio", "evidence": "At 4:00, the zebra says :'Hi, officer. Is there a problem?'", "inference": "He is in the polite tone."}, {"modality": "audio", "evidence": "At 4:03, the lion says hi.", "inference": "He is in the polite tone."}]}
{"id": "video_224__q0000", "video_id": "video_224", "qid": 0, "question_type": "ego reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_224_t5s.mp4"], "extra_info": {"question": "Why am I yelling?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because I overcame the difficulties.", "B. Because I saw the waterfall.", "C. Because I rolled in the water.", "D. Because I got through a narrow hole."], "correct_option": "A"}, "answer_text": "Because I overcame the difficulties.", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:16 in the video, 'I' was yelling.", "inference": "We need to rewind the video time to find out why 'I' was yelling."}, {"modality": "vision", "evidence": "At 3:55, 'I' saw the raging waterfall.", "inference": "It seemed difficult to kayak through such raging waterfall."}, {"modality": "vision", "evidence": "At 4:12, 'I' successfully navigated my kayak over the waterfall.", "inference": "'I' successfully overcame the difficulties by navigating over the raging waterfall, after which I yelled out at 4:16. Thus, it can be inferred that I was yelling because I overcame the difficulties."}]}
{"id": "video_228__q0000", "video_id": "video_228", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_228_t5s.mp4"], "extra_info": {"question": "When I say 'looking' what do I mean?"}, "reward_model_ground_truth": {"multi_choice": ["A. Look the wave.", "B. Look the sun.", "C. Look the rock.", "D. Look the beach."], "correct_option": "A"}, "answer_text": "Look the wave.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:13, I say 'Oh my god, looking!'", "inference": "This is the moment mentioned in the question."}, {"modality": "vision", "evidence": "At 0:11, I am knocked over by the waves.", "inference": "This indicates that the waves may have been too big."}, {"modality": "vision", "evidence": "At 0: 14, there are still big waves not far from me.", "inference": "The protagonist is showing the big waves to others."}]}
{"id": "video_230__q0000", "video_id": "video_230", "qid": 0, "question_type": "ego reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_230_t5s.mp4"], "extra_info": {"question": "When hearing the alarm clock ring, what can I see displayed on the TV in front of us?"}, "reward_model_ground_truth": {"multi_choice": ["A. 9:00!", "B. YOU'RE LATE!", "C. WAKE UP!", "D. GET UP!"], "correct_option": "C"}, "answer_text": "WAKE UP!", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:41, we can hear the alarm clock rings.", "inference": "To check what can I see on the TV."}, {"modality": "vision", "evidence": "At the same moment, 'WAKE UP!' is displayed on the TV in front of us.", "inference": "I can see 'WAKE UP!' on the TV when the alarm clock rings."}]}
{"id": "video_249__q0000", "video_id": "video_249", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_249_t5s.mp4"], "extra_info": {"question": "Which of the following is the correct order of the examples introducing food sovereignty?"}, "reward_model_ground_truth": {"multi_choice": ["A.hosts harvest festivals, the knowledge impartation by local chefs, the Cherokee nation the three sisters project", "B.the Cherokee nation, hosts harvest festivals, the knowledge impartation by local chefs, the three sisters project", "C.hosts harvest festivals, the Cherokee nation,  the three sisters project , the knowledge impartation by local chefs", "D.the Cherokee nation, the three sisters project, hosts harvest festivals, the knowledge impartation by local chefs"], "correct_option": "C"}, "answer_text": "hosts harvest festivals, the Cherokee nation, the three sisters project , the knowledge impartation by local chefs", "reasoning_steps": [{"modality": "vision", "evidence": "The video first introduces the Mongol Food Sovereignty Project, mentioning harvest festivals and community gardens.", "inference": "This is the initial example in the sequence, corresponding to 'hosts harvest festivals' among the options."}, {"modality": "vision", "evidence": "Next, the Cherokee Nation is mentioned for providing free heirloom seed packets.", "inference": "This acts as the second example, matching 'the Cherokee nation' in the options."}, {"modality": "audio", "evidence": "Then, the Three Sisters Project at Iowa State University is discussed.", "inference": "This serves as the third example, aligning with 'the three sisters project' in the options."}]}
{"id": "video_25__q0000", "video_id": "video_25", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_25_t5s.mp4"], "extra_info": {"question": "What are the characteristics of the little boys who mock the main characters?"}, "reward_model_ground_truth": {"multi_choice": ["A. There is a pot on his head.", "B. Two ponytails.", "C. Golden hair.", "D. Orange hair."], "correct_option": "D"}, "answer_text": "Orange hair.", "reasoning_steps": [{"modality": "audio", "evidence": "At 9:09, one boy is mocking.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "At 9:11, the little boys who mock the main characters has orange hair.", "inference": "Orange hair."}]}
{"id": "video_266__q0000", "video_id": "video_266", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_266_t5s.mp4"], "extra_info": {"question": "When I say 'oh my god', how many cars are in my field of vision??"}, "reward_model_ground_truth": {"multi_choice": ["A. 1.", "B. 2.", "C. 3.", "D. 4."], "correct_option": "B"}, "answer_text": "2.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:48, a man's voice says' Oh my god'.", "inference": "I say 'oh my god'."}, {"modality": "vision", "evidence": "At 2: 47, there is 1 car in front of this car, 2 cars in the opposite direction.", "inference": "There are totally 3 cars in my field of vision"}]}
{"id": "video_274__q0000", "video_id": "video_274", "qid": 0, "question_type": "ego reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_274_t5s.mp4"], "extra_info": {"question": "When hearing the opera singing from the earphones, what did I run past?"}, "reward_model_ground_truth": {"multi_choice": ["A. The vendor's cart.", "B. The bus stop.", "C. The subway station.", "D. The traffic light."], "correct_option": "B"}, "answer_text": "The bus stop.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:18, the opera began to sing.", "inference": "I heard the opera singing from the earphones."}, {"modality": "audio", "evidence": "At 6:20, the opera was still singing.", "inference": "I was hearing the opera singing from the earphones."}, {"modality": "vision", "evidence": "At 6:20, a man was running past a bus stop.", "inference": "When hearing the opera singing from the earphones, I run past the bus stop."}]}
{"id": "video_293__q0000", "video_id": "video_293", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_293_t5s.mp4"], "extra_info": {"question": "In the video, how many of the races commented on by the female commentator featured a red track?"}, "reward_model_ground_truth": {"multi_choice": ["A.1", "B.2", "C.3", "D.4"], "correct_option": "A"}, "answer_text": "1", "reasoning_steps": [{"modality": "audio", "evidence": "in this audio, only three of the races commented on by the female commentator.", "inference": "get the 3 races."}, {"modality": "vision", "evidence": "only one of the races commented on by the female commentator featured a red track.", "inference": "1."}]}
{"id": "video_298__q0000", "video_id": "video_298", "qid": 0, "question_type": "ego reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_298_t5s.mp4"], "extra_info": {"question": "What state was I in when the machine's sound rang out for the third time?"}, "reward_model_ground_truth": {"multi_choice": ["A.Holding sauce in hand.", "B.Pressing the sushi.", "C.Place shrimps on the sushi.", "D.Put shredded cucumbers on the sushi."], "correct_option": "C"}, "answer_text": "Place shrimps on the sushi.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:30, the machine makes its first sound.", "inference": "This is the first sound."}, {"modality": "vision", "evidence": "At 0:30, I was holding sauce in my hand.", "inference": "I was holding sauce."}, {"modality": "audio", "evidence": "At 0:51, the machine makes its second sound.", "inference": "This is the second sound."}, {"modality": "vision", "evidence": "At 0:51, I was pressing the sushi.", "inference": "I was pressing sushi."}, {"modality": "audio", "evidence": "At 1:52, the machine makes its third sound.", "inference": "This is the third sound."}, {"modality": "vision", "evidence": "At 1:52, I was placing shrimp on the sushi.", "inference": "I was placing shrimp on the sushi."}]}
{"id": "video_3__q0000", "video_id": "video_3", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_3_t5s.mp4"], "extra_info": {"question": "How many birds did the white dog scare away when the owner said, 'much for today'?"}, "reward_model_ground_truth": {"multi_choice": ["A.0", "B.1", "C.2", "D.3"], "correct_option": "C"}, "answer_text": "2", "reasoning_steps": [{"modality": "audio", "evidence": "The owner says 'Much for today' is at 2:04", "inference": "The dog will scare away the birds around this time."}, {"modality": "vision", "evidence": "There are two birds on the ground at 2:04.", "inference": "The dog will scare these two birds."}, {"modality": "vision", "evidence": "The two birds fly to the sky 2:05.", "inference": "The dog  scares away two birds."}]}
{"id": "video_342__q0000", "video_id": "video_342", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_342_t5s.mp4"], "extra_info": {"question": "What event immediately precedes the skier's shouts?"}, "reward_model_ground_truth": {"multi_choice": ["A. Navigating the final and narrowest turn.", "B. Safely reducing speed as the terrain flattens and widens.", "C. Successfully getting into the open water.", "D. Clearing the last rock obstacle in the gully."], "correct_option": "C"}, "answer_text": "Successfully getting into the open water.", "reasoning_steps": [{"modality": "audio", "evidence": "At 00:53, the skier begins shouting 'Yeah!' and 'Yes!'.", "inference": "The question asks what happens immediately after 00:53."}, {"modality": "vision", "evidence": "At 00:54, the skier exits the end of the ice channel.", "inference": "The skier exits the end of the ice channel after 00:53."}, {"modality": "vision", "evidence": "At 00:55, the skier enters a wide, open body of water.", "inference": "The skier successfully get into the open water after his shouts."}]}
{"id": "video_343__q0000", "video_id": "video_343", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_343_t5s.mp4"], "extra_info": {"question": "What was the immediate result of the shot that caused the Serbian female athlete to pump her fist?"}, "reward_model_ground_truth": {"multi_choice": ["A. It was the final, match-winning point.", "B. It set a new Olympic record.", "C. The score became tied.", "D. It caused the opposing team to call a timeout."], "correct_option": "C"}, "answer_text": "The score became tied.", "reasoning_steps": [{"modality": "vision", "evidence": "At 1:14, the Serbian female athlete was shown in the left pf the scene.", "inference": "Pay attention to her action."}, {"modality": "vision", "evidence": "At 1:21, the Serbian female athlete pumped her fist.", "inference": "It might be the result of the competition that made this, pay attention to this result."}, {"modality": "audio", "evidence": "At 1:18, the commentator says the score ' to level at 10-10'.", "inference": "This indicates that the two teams are tied in score. So the answer is 'The score became tied.'."}]}
{"id": "video_344__q0000", "video_id": "video_344", "qid": 0, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_344_t5s.mp4"], "extra_info": {"question": "When the driver throws a package, the sound it makes suggests its contents are most likely what?"}, "reward_model_ground_truth": {"multi_choice": ["A. Soft goods like clothing.", "B. Dense items like books.", "C. Fragile items.", "D. Glass."], "correct_option": "B"}, "answer_text": "Dense items like books.", "reasoning_steps": [{"modality": "vision", "evidence": "at the beginning of the video, the driver appears.", "inference": "get the man."}, {"modality": "vision", "evidence": "the driver throws a package at 0:16.", "inference": "locate this point."}, {"modality": "audio", "evidence": "upon landing, there is a loud sound of shattering.", "inference": "not so fragile, dense items like books."}]}
{"id": "video_344__q0001", "video_id": "video_344", "qid": 1, "question_type": "summarization", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_344_t5s.mp4"], "extra_info": {"question": "At the moment the dog ends barking, what is the delivery driver doing?"}, "reward_model_ground_truth": {"multi_choice": ["A. Ringing the doorbell.", "B. Walking back to his truck.", "C. Taking a delivery confirmation photo.", "D. Placing a package."], "correct_option": "B"}, "answer_text": "Walking back to his truck.", "reasoning_steps": [{"modality": "audio", "evidence": "a dog barks clearly at 0:08.", "inference": "locate the audio cue."}, {"modality": "vision", "evidence": "at 0:06, a delivery driver carrying a package is placing it", "inference": "get the man."}, {"modality": "vision", "evidence": "he is walking back to his truck.", "inference": "Walking back to his truck."}]}
{"id": "video_345__q0000", "video_id": "video_345", "qid": 0, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_345_t5s.mp4"], "extra_info": {"question": "Why does the man in the camouflage cap smile when being asked?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is happy to be helping a child.", "B. He finds the little girl's actions amusing.", "C. He recognizes the little girl as a well-known YouTuber", "D. He realizes it is a joke."], "correct_option": "D"}, "answer_text": "He realizes it is a joke.", "reasoning_steps": [{"modality": "vision", "evidence": "a girl is pretending buckets are extremely heavy.", "inference": "get the point."}, {"modality": "audio", "evidence": "there are comedic sound effects.", "inference": "this is a joke."}, {"modality": "vision", "evidence": "There are many adults here who can't afford to carry it, but little girls can.", "inference": "this is just a joke."}, {"modality": "vision", "evidence": "the man smiles before even touching the buckets.", "inference": "He realizes it is a joke."}]}
{"id": "video_345__q0001", "video_id": "video_345", "qid": 1, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_345_t5s.mp4"], "extra_info": {"question": "After the sound of coins is heard, how many of the people who subsequently attempt to lift the buckets are wearing glasses?"}, "reward_model_ground_truth": {"multi_choice": ["A. Zero.", "B. One.", "C. Two.", "D. Three."], "correct_option": "C"}, "answer_text": "Two.", "reasoning_steps": [{"modality": "audio", "evidence": "the sound of coins is heard around 0:09.", "inference": "locate the audio cue."}, {"modality": "vision", "evidence": "the first man attempts to lift the buckets.", "inference": "no glasses."}, {"modality": "vision", "evidence": "the second woman attempts to lift the buckets.", "inference": "no glasses."}, {"modality": "vision", "evidence": "the third woman attempts to lift the buckets.", "inference": "no glasses."}, {"modality": "vision", "evidence": "the fourth man attempts to lift the buckets.", "inference": "wear glasses."}, {"modality": "vision", "evidence": "the fifth man attempts to lift the buckets.", "inference": "there are 2 men, who wear glasses."}]}
{"id": "video_346__q0000", "video_id": "video_346", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_346_t5s.mp4"], "extra_info": {"question": "What is Carlos Alcaraz doing at the exact moment the commentator mentions 'David Ferrer'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Toweling off near his chair.", "B. Talking to the umpire.", "C. Walking towards the baseline.", "D. Looking at his coaching box."], "correct_option": "C"}, "answer_text": "Walking towards the baseline.", "reasoning_steps": [{"modality": "audio", "evidence": "At 00:17 in the video, the commentator mentions 'David Ferrer'.", "inference": "We know 'the exact moment' mentioned in the question should be around 00:17."}, {"modality": "vision", "evidence": "At 00:17 in the video, we can see the face of Carlos Alcaraz.", "inference": "As the question suggests, it asks what this man is doing at about 00:17."}, {"modality": "vision", "evidence": "At 00:18 in the video, Carlos Alcaraz is walking towards the baseline for the next point.", "inference": "The answer is walking towards the baseline."}]}
{"id": "video_347__q0000", "video_id": "video_347", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_347_t5s.mp4"], "extra_info": {"question": "What are the characteristics of the clip when the host mentions that the goal is similar to a shot by a core player from the NBA's 2011 championship team?"}, "reward_model_ground_truth": {"multi_choice": ["A.The scoring player's name is Wang Zhe lin.", "B.The scoring player's jersey number is 10.", "C.The New Zealand team has committed more than one foul.", "D.At this moment, player number 0 is under the basket, ready to inbound the ball."], "correct_option": "B"}, "answer_text": "The scoring player's jersey number is 10.", "reasoning_steps": [{"modality": "audio", "evidence": "At 00:50, the host mentions, 'This is Dirk Nowitzki's shot!'", "inference": "As we know Dirk Nowitzki was 'a core player from the NBA's 2011 championship team' as mentioned in the question, the question might be asking the characteristics of the clip at around 00:50."}, {"modality": "vision", "evidence": "At 00:48, a player scores with a shot.", "inference": "A player scores when the host mentions 'Dirk Nowitzki's shot'. And we can see what he looks like."}, {"modality": "vision", "evidence": "At 00:46, we can see the player's jersey number is 10.", "inference": "The answer should be 'The scoring player's jersey number is 10'."}]}
{"id": "video_348__q0000", "video_id": "video_348", "qid": 0, "question_type": "background&music understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_348_t5s.mp4"], "extra_info": {"question": "In the video, which of the following game sound effects sounds the most distinctive? Digging dirt, digging stone, placing wooden planks, placing chests."}, "reward_model_ground_truth": {"multi_choice": ["A.Digging dirt", "B.Digging stone", "C.Placing wooden planks", "D.Placing chests"], "correct_option": "A"}, "answer_text": "Digging dirt", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:01, the character in the video is placing wooden planks.", "inference": "The sound of 'placing wooden planks' is mentioned in the question."}, {"modality": "audio", "evidence": "At 00:01, we can hear a short, muffled 'thud' sound of medium intensity.", "inference": "The sound of placing wooden planks is a short 'thud' sound of medium intensity."}, {"modality": "vision", "evidence": "At 00:09, the character is digging dirt.", "inference": "The sound of 'digging dirt' is mentioned in the question."}, {"modality": "audio", "evidence": "At 00:09, we can hear a crisp crackling sound that is similar to chewing potato chips.", "inference": "The sound of digging dirt planks is different from the sound of placing wooden planks. It's a crisp crackling sound."}, {"modality": "vision", "evidence": "At 00:30, the character in the video is digging stone.", "inference": "The sound of 'digging stone' is mentioned in the question."}, {"modality": "audio", "evidence": "At 00:30, we can hear a lighter, shorter, muffled 'thud' sound.", "inference": "The sound of digging stone is a lighter 'thud' sound. It is similar to the sound of placing wooden planks, but it is lighter."}, {"modality": "vision", "evidence": "At 00:44, the character in the video is placing chests.", "inference": "The sound of 'placing chests' is mentioned in the question."}, {"modality": "audio", "evidence": "At 00:44, we can hear a short, muffled 'thud' sound of medium intensity.", "inference": "The sound of placing chests is very similar to the sound of placing wooden planks. In conclusion, the sounds of placing chests, digging stone, and placing wooden planks are highly similar to one another, while the sound of digging dirt is quite different from the other three. So it can be inferred that the sound of digging dirt sounds the most distinctive."}]}
{"id": "video_349__q0000", "video_id": "video_349", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_349_t5s.mp4"], "extra_info": {"question": "Who was the last shot for Shanghai Port in the first half in this video?"}, "reward_model_ground_truth": {"multi_choice": ["A.Zhu Chenjie", "B.Gao Tianyi", "C.Liu Ruofan", "D.Leonardo"], "correct_option": "C"}, "answer_text": "Liu Ruofan", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:04, the commentator says: Liu Ruofan rushes into the penalty area!", "inference": "To check if this player is from Shanghai Port"}, {"modality": "vision", "evidence": "At  2:05, Liu Ruofan in Shanghai Port's red uniform made a shot in the first half.", "inference": "To check if this is the last shot in the first half."}, {"modality": "vision", "evidence": "At 2:25,  a shot contributed by Leonardo from Shanghai Port in the second half.", "inference": "Liu Ruofan made the last shot for Shanghai Port in the first half."}]}
{"id": "video_35__q0000", "video_id": "video_35", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_35_t5s.mp4"], "extra_info": {"question": "What is the first sentence Luisa said in this video?"}, "reward_model_ground_truth": {"multi_choice": ["A. What's with the no-show at the meeting.", "B. I bet it's in my spam.", "C. Oh, you guys are taking the piss", "D. Are you sitting the meeting out?"], "correct_option": "C"}, "answer_text": "Oh, you guys are taking the piss", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:54, a woman was drinking almond milk.", "inference": "Identify the almond milk."}, {"modality": "vision", "evidence": "At 2:54, there was a blue label attached to the carton of almond milk.", "inference": "It will most likely have the name of the owner of this carton of almond milk written on it."}, {"modality": "vision", "evidence": "The blue label was written 'Luisa'.", "inference": "The owner of this carton of almond milk is Luisa."}, {"modality": "vision", "evidence": "At 3:52, A woman was holding this carton of almond milk in anger.", "inference": "She is Luisa, pay attention to the first sentence she said."}, {"modality": "audio", "evidence": "At 3:47, Luisa said 'Oh, you guys are taking the piss'.", "inference": "So the answer is 'Oh, you guys are taking the piss'."}]}
{"id": "video_350__q0000", "video_id": "video_350", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_350_t5s.mp4"], "extra_info": {"question": "When the players of both teams enter the field, according to what the commentator says, why today's match is particularly worth celebrating?"}, "reward_model_ground_truth": {"multi_choice": ["A.Team in black is celebrating its 125th anniversary.", "B.Team in pink is celebrating its 125th anniversary.", "C.Team in black is celebrating its 126th anniversary.", "D.Team in pink is celebrating its 126th anniversary."], "correct_option": "B"}, "answer_text": "Team in pink is celebrating its 125th anniversary.", "reasoning_steps": [{"modality": "vision", "evidence": "at the beginning,the players enter the field.", "inference": "get the point."}, {"modality": "audio", "evidence": "then, the narrator says: 'And City have the opportunity to join the celebrations of the 125th birthday of this famous old club.'", "inference": "Man City has the honor to join in celebrating the birthday of another football club."}, {"modality": "vision", "evidence": "the visuals show that today's match is Man City against Palermo.", "inference": "the club celebrating its 125th anniversary is Palermo."}, {"modality": "audio", "evidence": "the narrator introduces the Palermo team, who are wearing pink jerseys.", "inference": "Team in pink is celebrating its 125th anniversary."}]}
{"id": "video_350__q0001", "video_id": "video_350", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_350_t5s.mp4"], "extra_info": {"question": "How many players wearing pink jerseys are visible in the frame when the referee ends the match with the whistle?"}, "reward_model_ground_truth": {"multi_choice": ["A.7.", "B.6.", "C.9.", "D.10."], "correct_option": "D"}, "answer_text": "10", "reasoning_steps": [{"modality": "audio", "evidence": "Around 8:35 in the video, the referee blows the whistle three times in succession", "inference": "locate this point."}, {"modality": "audio", "evidence": "the commentator says: 'he is going to stops his time. ...   looks at his watch, and that will do.   And City has won by 3 goals.'", "inference": "At this point, the match ends."}, {"modality": "vision", "evidence": "during this time,10 players wearing pink jerseys can be clearly seen.", "inference": "10"}]}
{"id": "video_351__q0000", "video_id": "video_351", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_351_t5s.mp4"], "extra_info": {"question": "What were the results of Liverpool's 5 penalties in the Penalty Shootout?"}, "reward_model_ground_truth": {"multi_choice": ["A.Saved miss, saved miss, goal, saved miss, goal.", "B.Off target miss, saved miss, goal, saved miss, goal.", "C.Off target miss, goal, goal, saved miss, off target miss.", "D.Goal, saved miss, goal, saved miss, goal."], "correct_option": "B"}, "answer_text": "Off target miss, saved miss, goal, saved miss, goal.", "reasoning_steps": [{"modality": "audio", "evidence": "at 3:06, the man said' Liverpool is looking for...'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "the man in white is from  Liverpool team.", "inference": "the players in white is from Liverpool team."}, {"modality": "vision", "evidence": "at 3:06, Liverpool's first penalty flies over the crossbar, missing the target,", "inference": "Off target miss."}, {"modality": "vision", "evidence": "Liverpool's second penalty is saved by Crystal Palace'sgoalkeeper Henderson with his left hand.", "inference": "saved miss."}, {"modality": "vision", "evidence": "Liverpool's third penalty hits the target.", "inference": "goal."}, {"modality": "vision", "evidence": "Liverpool's fourth penalty is saved by Crystal Palace's goalkeeper.", "inference": "saved miss."}, {"modality": "vision", "evidence": "Liverpool's fifth penalty hits the target.", "inference": "goal."}]}
{"id": "video_352__q0000", "video_id": "video_352", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_352_t5s.mp4"], "extra_info": {"question": "After the blogger said 'rag is an AI workflow,' which company's logo is the third icon in the first row of the six toolkits that appeared?"}, "reward_model_ground_truth": {"multi_choice": ["A.Openai", "B.Amazon", "C.Cohere", "D.Perplexity"], "correct_option": "D"}, "answer_text": "Perplexity", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:05, the man said 'rag is an AI workflow'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "there are 6 toolkits on the screen.", "inference": "we should pay attention to the third icon in the first row of the six toolkits that appeared."}, {"modality": "vision", "evidence": "the third icon in the first row of the six toolkits is Perplexity.", "inference": "Perplexity"}]}
{"id": "video_352__q0001", "video_id": "video_352", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_352_t5s.mp4"], "extra_info": {"question": "When a blogger presents the process of searching for a skier, which of the following is correct?"}, "reward_model_ground_truth": {"multi_choice": ["A.The timestamp of the image extracted from the video is 1:21.", "B.There were 4 clips found that were not highly relevant.", "C.The length of the second viewed clip is 8 seconds.", "D.The fifth clip in the search results has a duration of 8 seconds."], "correct_option": "B"}, "answer_text": "There were 4 clips found that were not highly relevant.", "reasoning_steps": [{"modality": "audio", "evidence": "The blogger said he will search for a skier.", "inference": "locate this point."}, {"modality": "vision", "evidence": "The blogger is showing the search for a skier.", "inference": "locate this point."}, {"modality": "vision", "evidence": "There are 10 clips searched overall.", "inference": "locate this point."}, {"modality": "vision", "evidence": "there are 6 clips highly relevant with the video.", "inference": "there are 4 videos found that were not highly relevant."}]}
{"id": "video_354__q0000", "video_id": "video_354", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_354_t5s.mp4"], "extra_info": {"question": "How much experience is needed to research a Tier XI tank?"}, "reward_model_ground_truth": {"multi_choice": ["A.325000", "B.90000", "C.7400000", "D.273500"], "correct_option": "A"}, "answer_text": "325000", "reasoning_steps": [{"modality": "audio", "evidence": "the voiceover introduces tier 11 tanks in 3:36", "inference": "narrow the range"}, {"modality": "vision", "evidence": "At timestamp 03:56, the video displays the German tech tree.", "inference": "experience needed to research a Tier XI tank can be found in it"}, {"modality": "vision", "evidence": "the 'Taschenratte' is shown with a research cost of 325,000 experience points.", "inference": "the answer is 325000"}]}
{"id": "video_354__q0001", "video_id": "video_354", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_354_t5s.mp4"], "extra_info": {"question": "Where are the vehicle's crews in the garage interface?"}, "reward_model_ground_truth": {"multi_choice": ["A.The underside of the main vehicle", "B.The left of the screen", "C.Directly to the left of the ammunition bar", "D.Directly to the right of the equipment"], "correct_option": "A"}, "answer_text": "The underside of the main vehicle", "reasoning_steps": [{"modality": "audio", "evidence": "the voiceover introduces the new garage interface at 01:26.", "inference": "narrow the range"}, {"modality": "vision", "evidence": "the garage interface is shown clearly starting at 01:26.", "inference": "the vehicle's crews is in the garage interface"}, {"modality": "vision", "evidence": "The  crew members lay out is under the main vehicle on the left side", "inference": "the answer is 'The underside of the main vehicle'."}]}
{"id": "video_355__q0000", "video_id": "video_355", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_355_t5s.mp4"], "extra_info": {"question": "When the singer first escalates her anger from vocal expression to physical contact with another singer, how does the orchestra musically intensify the dramatic conflict?"}, "reward_model_ground_truth": {"multi_choice": ["A. The entire orchestra plays a single, extremely loud, short chord.", "B. The music suddenly stops completely, creating a breathtaking moment.", "C. The strings play a series of rapidly descending notes, conveying the character's sense of fall.", "D. Only the brass instruments play a single, piercing, high note."], "correct_option": "A"}, "answer_text": "The entire orchestra plays a single, extremely loud, short chord.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:05, the singer starts signing with anger.", "inference": "Identify the signer."}, {"modality": "vision", "evidence": "At 0:05, the singer wears a black dress.", "inference": "Pay attention to the time point when she makes physical contact with another singer."}, {"modality": "vision", "evidence": "At 0:24, this singer in a black dress is making physical contact with another singer.", "inference": "Note what kind of music the orchestra is depicting at this moment."}, {"modality": "audio", "evidence": "At 0:24, the orchestra lets out a single, extremely loud, short chord.", "inference": "So the answer is 'The entire orchestra plays a single, extremely loud, short chord.'."}]}
{"id": "video_356__q0000", "video_id": "video_356", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_356_t5s.mp4"], "extra_info": {"question": "In this video, how many parts do the two sing together?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 5"], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision", "evidence": "At the beginning of the video, there are two spliced videos in the scene.", "inference": "This is most likely a collaborative performance by the two on the left and right."}, {"modality": "vision", "evidence": "The left side of the scene shows an animated cat.", "inference": "It is one of the singers."}, {"modality": "vision", "evidence": "The right side of the scene shows the NetEase Cloud Music cover for one of Shen Jiarun's songs.", "inference": "It is the other singers, pay attention to the part where they perform a chorus together afterwards."}, {"modality": "audio", "evidence": "At 0:10, they sing together.", "inference": "This is the first time."}, {"modality": "audio", "evidence": "At 0:17, they sing together.", "inference": "This is the second time."}, {"modality": "audio", "evidence": "At 0:47, they sing together.", "inference": "This is the third time."}, {"modality": "audio", "evidence": "At 0:55, they sing together.", "inference": "This is the fourth time. So the answer is '4'."}]}
{"id": "video_357__q0000", "video_id": "video_357", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_357_t5s.mp4"], "extra_info": {"question": "In the first song, which level is not on the list?"}, "reward_model_ground_truth": {"multi_choice": ["A. Balloon Jump", "B. Side Flip", "C. Kick the Hat on Your Head", "D. Chop the Balloon"], "correct_option": "B"}, "answer_text": "Side Flip", "reasoning_steps": [{"modality": "audio", "evidence": "The first song at the start of the video begins to play immediately.", "inference": "Pay attention to what actions the people in the video are doing during the first song."}, {"modality": "vision", "evidence": "At 0:04, a man is making a 'Balloon Jump'.", "inference": "This is the first level."}, {"modality": "vision", "evidence": "At 0:15, a man is making a 'Chop the Balloon'.", "inference": "This is the second level."}, {"modality": "vision", "evidence": "At 0:27, a man is making a 'Kick the Hat on Your Head'.", "inference": "This is the third level, so the level of 'Side Flip' is not in the list."}]}
{"id": "video_358__q0000", "video_id": "video_358", "qid": 0, "question_type": "reference reasoning", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_358_t5s.mp4"], "extra_info": {"question": "Who is the second main dancer of the first song in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. The girl in yellow", "B. The girl in red", "C. The girl in pink", "D. The girl in white"], "correct_option": "C"}, "answer_text": "The girl in pink", "reasoning_steps": [{"modality": "audio", "evidence": "At the beginning of the video, the song starts.", "inference": "Pay attention to who is standing at the center of the team, she is the first main dancer."}, {"modality": "vision", "evidence": "At 0:03, the girl in red takes center stage.", "inference": "She is the first main dancer."}, {"modality": "vision", "evidence": "At 0:06, the girl in pink takes center stage.", "inference": "She is the second lead dancer, so the answer is 'The girl in pink'."}]}
{"id": "video_359__q0000", "video_id": "video_359", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_359_t5s.mp4"], "extra_info": {"question": "How many octaves do the three instruments span when tuned in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "B"}, "answer_text": "2", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:01, subtitles appear in the scene says 'It's bad enough that the people upstairs is playing the violin in the middle of the night.'.", "inference": "The first instrument is the violin."}, {"modality": "vision", "evidence": "At 0:03, subtitles appear in the scene says 'And someone downstairs is also playing the cello.'.", "inference": "The second instrument is the cello."}, {"modality": "vision", "evidence": "At 0:09, a piano appears in the scene.", "inference": "This indicates that the person filming the video is about to play the piano, so the piano is the third instrument."}, {"modality": "audio", "evidence": "At 0:28, the piano is at D4.", "inference": "This is the first octave."}, {"modality": "audio", "evidence": "At 0:31, the cello is at D3.", "inference": "This is the second octave. So the answer is '2'."}]}
{"id": "video_36__q0000", "video_id": "video_36", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_36_t5s.mp4"], "extra_info": {"question": "Four new people joined this hospital today, when a doctor asked 'what might made her faint on the plat?', who are the new joined people that answered this question?"}, "reward_model_ground_truth": {"multi_choice": ["A. Melissa Kim and Victoria Javadi", "B. Melissa Kim and Trinity Santos", "C. Victoria Javadi and Denise Whitaker", "D. Trinity Santos and Denise Whitaker"], "correct_option": "C"}, "answer_text": "Victoria Javadi and Denise Whitaker", "reasoning_steps": [{"modality": "audio", "evidence": "At 01:48, a male doctor introduced a female doctor, 'starting with second year resident Dr Melissa King'.", "inference": "Dr Melissa King is one of the new people joined this hospital today."}, {"modality": "audio", "evidence": "At 01:58, a girl introduced herself 'Trinity Santos, intern'.", "inference": "Trinity Santos is also one of the new people joined this hospital today."}, {"modality": "audio", "evidence": "At 01:59, a girl introduced herself 'Victoria Javadi, MS3'.", "inference": "Victoria Javadi is also one of the new people joined this hospital today."}, {"modality": "audio", "evidence": "At 02:00, a boy introduced himself 'Dennis Whitaker, MS4'.", "inference": "Dennis Whitaker is also one of the new people joined this hospital today."}, {"modality": "audio", "evidence": "At 05:52, a male doctor asked, 'what might have made her faint on the platform?'", "inference": "identify the question."}, {"modality": "audio", "evidence": "At 05:52, a girl's voice and a boy's voice answered the question.", "inference": "so they are a girl and Denise Whitaker."}, {"modality": "vision", "evidence": "At 05:52, the girl who answered the question is Victoria Javadi.", "inference": "so they are Victoria Javadi and Denise Whitaker."}]}
{"id": "video_360__q0000", "video_id": "video_360", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_360_t5s.mp4"], "extra_info": {"question": "In the video, how many times did the dog and the person play the pitch together?"}, "reward_model_ground_truth": {"multi_choice": ["A. 3", "B. 4", "C. 5", "D. 6"], "correct_option": "A"}, "answer_text": "3", "reasoning_steps": [{"modality": "vision", "evidence": "at the beginning, there are a man, a dog and a cat.", "inference": "get the man and the dog."}, {"modality": "audio", "evidence": "the man is playing the piano.", "inference": "get the point."}, {"modality": "vision", "evidence": "at 0:13, the man points a piano key.", "inference": "maybe the dog will press it."}, {"modality": "vision", "evidence": "at 0:13, the dog pressed it.", "inference": "this is the first time that the dog and the person play the pitch together."}, {"modality": "vision", "evidence": "at 0:14, the dog pressed it.", "inference": "this is the second time that the dog and the person play the pitch together."}, {"modality": "vision", "evidence": "at 0:16, the dog pressed it again.", "inference": "the total is 3."}]}
{"id": "video_361__q0000", "video_id": "video_361", "qid": 0, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_361_t5s.mp4"], "extra_info": {"question": "Why is the man in black laughing?"}, "reward_model_ground_truth": {"multi_choice": ["A. The trumpet player looks funny.", "B. The trumpet player is about to fall down because he's not sitting firmly.", "C. The trumpet player and the chair are playing in harmony.", "D. He was born to laugh."], "correct_option": "C"}, "answer_text": "The trumpet player and the chair are playing in harmony.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:01, a trumpeter is playing the trumpet.", "inference": "The trumpeter plays the trumpet."}, {"modality": "vision", "evidence": "At 00:02, the trumpeter moved the chair.", "inference": "The trumpeter moves the chair."}, {"modality": "audio", "evidence": "At 0:02, the chair produced a big sound.", "inference": "The chair produces a big sound while the trumpeter is playing the trumpet. It sounds like they play in harmony."}, {"modality": "audio", "evidence": "At 0:02, someone starts laughing.", "inference": "This should be the laughing man in the question."}, {"modality": "vision", "evidence": "At 0:02, we can see the one laughing is the man in black.", "inference": "The man in black is laughing when the chair produces a big sound. In conclusion, it can be inferred that he's laughing because the trumpet player and the chair are playing in harmony."}]}
{"id": "video_362__q0000", "video_id": "video_362", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_362_t5s.mp4"], "extra_info": {"question": "When adding which rhythmic pattern does it begin to resemble crosstalk?"}, "reward_model_ground_truth": {"multi_choice": ["A. Two eighth notes", "B. One quarter note", "C. One eighth note", "D. One whole rest"], "correct_option": "C"}, "answer_text": "One eighth note", "reasoning_steps": [{"modality": "audio", "evidence": "There are only quarter notes before 0:10.", "inference": "It doesn't  begin to resemble crosstalk."}, {"modality": "audio", "evidence": "Two eighth notes notes are adding after 0:10.", "inference": "It still doesn't  begin to resemble crosstalk."}, {"modality": "audio", "evidence": "It begins to resemble crosstalk at 0:26.", "inference": "Locate the time."}, {"modality": "vision", "evidence": "A rest and an one eighth note is adding at 0:25.", "inference": "The addition of an eighth note creates a rhythmic, irregular feel typical of crosstalk."}]}
{"id": "video_363__q0000", "video_id": "video_363", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_363_t5s.mp4"], "extra_info": {"question": "When everyone started playing, how many times did the beat in the video hit?"}, "reward_model_ground_truth": {"multi_choice": ["A. 17 times", "B. 16 times", "C. 13 times", "D. 11 times"], "correct_option": "C"}, "answer_text": "13 times", "reasoning_steps": [{"modality": "vision", "evidence": "At  0:14, everyone picks up their instruments and ready to play.", "inference": "Set the point to count the beat."}, {"modality": "audio", "evidence": "At  0:15, their playing matches the beat for the 1st time.", "inference": "The beat hits 1 times so far."}, {"modality": "audio", "evidence": "Also at 0:15, their playing matches the beat for the 2ndt time.", "inference": "The beat hits 2 times so far."}, {"modality": "audio", "evidence": "At 0:16, their playing matches the beat for the 3rd time.", "inference": "The beat hits 3 times so far."}, {"modality": "audio", "evidence": "Also at 0:16, their playing matches the beat for the 4th time.", "inference": "The beat hits 4 times so far."}, {"modality": "audio", "evidence": "At 0:17, their playing matches the beat for the 5th time.", "inference": "The beat hits 5 times so far."}, {"modality": "audio", "evidence": "At 0:18, their playing matches the beat for the 6th time.", "inference": "The beat hits 6 times so far."}, {"modality": "audio", "evidence": "Also at 0:18, their playing matches the beat for the 7th time.", "inference": "The beat hits 7 times so far."}, {"modality": "audio", "evidence": "At 0:19, their playing matches the beat for the 8th time.", "inference": "The beat hits 8 times so far."}, {"modality": "audio", "evidence": "Also at 0:19, their playing matches the beat for the 9th time.", "inference": "The beat hits 9 times so far."}, {"modality": "audio", "evidence": "At 0:20, their playing matches the beat for the 10th time.", "inference": "The beat hits 10 times so far."}, {"modality": "audio", "evidence": "At 0:21, their playing matches the beat for the 11th time.", "inference": "The beat hits 11 times so far."}, {"modality": "audio", "evidence": "At 0:21, their playing matches the beat for the 12th time.", "inference": "The beat hits 12 times so far."}, {"modality": "audio", "evidence": "At 0:22, their playing matches the beat for the 13th time.", "inference": "The beat hits 13 times in total."}]}
{"id": "video_365__q0000", "video_id": "video_365", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_365_t5s.mp4"], "extra_info": {"question": "How many pairs of rhythmic patterns in the video sound exactly the same?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 1"], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision", "evidence": "we can see that these pairs of rhythmic patterns are played by switching between the left and right hands.", "inference": "locate this point."}, {"modality": "vision", "evidence": "there are a total of nine rows of rhythms.", "inference": "locate this point."}, {"modality": "audio", "evidence": "through sound recognition, the rhythm of the second row is the same as that of the third row.", "inference": "this is the first pair."}, {"modality": "audio", "evidence": "through sound recognition, the rhythm of the fourth row is the same as that of the fifth row.", "inference": "this is the second pair."}, {"modality": "audio", "evidence": "through sound recognition, the rhythm of the sixth row is the same as that of the seventh row.", "inference": "this is the third pair."}, {"modality": "audio", "evidence": "through sound recognition, the rhythm of the eighth row is the same as that of the ninth row.", "inference": "the total is 4."}]}
{"id": "video_366__q0000", "video_id": "video_366", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_366_t5s.mp4"], "extra_info": {"question": "What moves do the dancers make when they hear the single hop for the fourth time?"}, "reward_model_ground_truth": {"multi_choice": ["A. Tap to the left", "B. Tap to the right", "C. Moonwalk to the left", "D. Moonwalk to the right"], "correct_option": "D"}, "answer_text": "Moonwalk to the right", "reasoning_steps": [{"modality": "audio", "evidence": "At the very beginning of the audio, a male voice in the background says 'one Hop this time'.", "inference": "This is the first time."}, {"modality": "audio", "evidence": "At 0:02, a male voice in the background says again 'one Hop this time'.", "inference": "This is the second time."}, {"modality": "audio", "evidence": "At 0:08, a male voice in the background says 'Slide to the left.'.", "inference": "This is the third time."}, {"modality": "audio", "evidence": "At 0:11, a male voice in the background says 'Slide to the right.'.", "inference": "This is the fourth time, pay attention to what moves do the dancers make at this time."}, {"modality": "vision", "evidence": "At 0:11, the dancers make a moonwalk to the right.", "inference": "So the answer is 'Moonwalk to the right'."}]}
{"id": "video_368__q0000", "video_id": "video_368", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_368_t5s.mp4"], "extra_info": {"question": "How many chords does the guitarist play in total?"}, "reward_model_ground_truth": {"multi_choice": ["A. 9", "B. 10", "C. 11", "D. 12"], "correct_option": "C"}, "answer_text": "11", "reasoning_steps": [{"modality": "audio", "evidence": "We can hear the guitarist plays the chord called E.", "inference": "1 chord played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called #Gm7.", "inference": "2 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays another chord called #Gm7.", "inference": "3 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called B.", "inference": "4 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called A.", "inference": "5 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called D.", "inference": "6 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called #Fm7.", "inference": "7 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called Am7,.", "inference": "8 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called B7.", "inference": "9 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called #A#.", "inference": "10 chords played so far."}, {"modality": "audio", "evidence": "We can hear the guitarist plays the chord called E7.", "inference": "11 chords played in total."}]}
{"id": "video_369__q0000", "video_id": "video_369", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_369_t5s.mp4"], "extra_info": {"question": "Which instrumentalist here was almost slacking off the entire time?"}, "reward_model_ground_truth": {"multi_choice": ["A. Keyboardist", "B. Singer", "C. Bassist", "D. Drummer"], "correct_option": "C"}, "answer_text": "Bassist", "reasoning_steps": [{"modality": "audio", "evidence": "The singer kept singing until 0:10.", "inference": "The singer is putting in effort to sing."}, {"modality": "vision", "evidence": "The drummer kept beating the rhythm until the end.", "inference": "The drummer is putting in effort too."}, {"modality": "audio", "evidence": "At 0:11, the keyboardist started knocking keyboard.", "inference": "The keyboardist is putting in effort too."}, {"modality": "vision", "evidence": "At 0:05, the Bassist behind was drinking cola.", "inference": "The Bassist was slacking at the beginning."}, {"modality": "audio", "evidence": "At 0:13, the Bassist only made a sound 'Ah'.", "inference": "The bassist isn't doing his job and he is lmost slacking off the entire time. So the answer is 'Bassist'."}]}
{"id": "video_37__q0000", "video_id": "video_37", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_37_t5s.mp4"], "extra_info": {"question": "There is a doctor who might go to contact a patient's family. The patient is scheduled for a test—how many hours later might the test be performed if the patient's pupils are dilated?"}, "reward_model_ground_truth": {"multi_choice": ["A. Two hours", "B. Three hours", "C. Four hours", "D. Six hours"], "correct_option": "A"}, "answer_text": "Two hours", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:16, a woman says 'Repeat head CT in three hours.'.", "inference": "It indicates that a patient might be scheduled for a head CT in three hours."}, {"modality": "audio", "evidence": "At 2:20, a woman continously says 'Or sooner, if he blows a pupil.'.", "inference": "It indicates that if this patient has dilated pupils, a head CT should be performed in advance that less of three hours. So the answer is 'Two hours'."}, {"modality": "audio", "evidence": "At 2:33, a woman said 'Is anyone notified the family? '", "inference": "She may be the doctor who might go to contact a patient's family."}, {"modality": "vision", "evidence": "At 2:33, a doctor with blode ponytail is shown in the scene.", "inference": "Identify her."}, {"modality": "audio", "evidence": "She said 'I can try.'.", "inference": "She might go to contact a patient's family."}]}
{"id": "video_370__q0000", "video_id": "video_370", "qid": 0, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_370_t5s.mp4"], "extra_info": {"question": "Why does this video evoke a sense of peace and prosperity?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because the lights and the song are both beautiful.", "B. The song's beautiful meaning matches the visuals.", "C. This song evokes a sense of peace and prosperity.", "D. This visual evokes a sense of peace"], "correct_option": "B"}, "answer_text": "The song's beautiful meaning matches the visuals.", "reasoning_steps": [{"modality": "audio", "evidence": "The melody of the background song is called 'China in Lights' .", "inference": "This is the song that symbolizes China's prosperity and beauty. This could be the feeling the shooter intends to express."}, {"modality": "vision", "evidence": "The scene is shot from an aerial perspective of an airplane.", "inference": "Pay attention to the scene presented at this moment."}, {"modality": "vision", "evidence": "The scene shows the splendid view of countless glowing homes throughout China.", "inference": "This scene matches the music that the shooter has paired with it perfectly. So the answer is 'The song's beautiful meaning matches the visuals.'."}]}
{"id": "video_371__q0000", "video_id": "video_371", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_371_t5s.mp4"], "extra_info": {"question": "When the child begins to overcome the obstacles, what instrument is added?"}, "reward_model_ground_truth": {"multi_choice": ["A. Voice", "B. Harpsichord", "C. Piano", "D. Drums"], "correct_option": "D"}, "answer_text": "Drums", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:04, there is the sound of harpsichord in the background music.", "inference": "Pay attention to what musical instrument is added to the background music after the boy begins to overcome the obstacles."}, {"modality": "vision", "evidence": "At 0:05, the child begins to overcome the obstacles.", "inference": "This is the moment mentioned in the question."}, {"modality": "audio", "evidence": "At 0:06, the drum beat joins into the background music.", "inference": "So the answer is 'Drums'."}]}
{"id": "video_375__q0000", "video_id": "video_375", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_375_t5s.mp4"], "extra_info": {"question": "When the kitten expresses fear, what instrument is used in the background?"}, "reward_model_ground_truth": {"multi_choice": ["A. Synthesizer", "B. Wind instrument", "C. String instrument", "D. Percussion instrument"], "correct_option": "C"}, "answer_text": "String instrument", "reasoning_steps": [{"modality": "vision", "evidence": "at 0:06, the cat shows its fear.", "inference": "locate this point."}, {"modality": "audio", "evidence": "a string instrument sound was heard.", "inference": "String instrument is used in the background."}]}
{"id": "video_376__q0000", "video_id": "video_376", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_376_t5s.mp4"], "extra_info": {"question": "The sound effect added here to represent the panda's cheerfulness resembles the sound of which animal?"}, "reward_model_ground_truth": {"multi_choice": ["A. Ostrich", "B. Chicken", "C. Cat", "D. Duck"], "correct_option": "D"}, "answer_text": "Duck", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:03, the panda started to run around cheerfully.", "inference": "The video reflected the panda's cheerfulness at that moment."}, {"modality": "audio", "evidence": "At 00:04, the sound effect was added.", "inference": "The sound effect was added to represent the panda's cheerfulness."}, {"modality": "audio", "evidence": "At 00:04, the sound effect was added.", "inference": "The sound was like that of a duck quacking, so the answer should be 'Duck'."}]}
{"id": "video_377__q0000", "video_id": "video_377", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_377_t5s.mp4"], "extra_info": {"question": "About two minutes into the video, how many choral parts are singing?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "B"}, "answer_text": "2", "reasoning_steps": [{"modality": "vision", "evidence": "At 02:00, we can see many choristers in the video.", "inference": "There are many choristers in the video."}, {"modality": "vision", "evidence": "At 02:00, some choristers open their mouths very wide and round.", "inference": "They might be singing higher notes."}, {"modality": "vision", "evidence": "At 02:00, some choristers shape their mouths into a flatter form.", "inference": "They might be singing lower notes."}, {"modality": "audio", "evidence": "At 02:00, we can hear someone singing at a higher pitch.", "inference": "There is at least one choral part singing at 2:00."}, {"modality": "audio", "evidence": "At 02:00, we can hear someone singing at a lower pitch.", "inference": "This is the second choral part at 2:00. In total, there should be two choral parts singing at 2:00 in the video."}]}
{"id": "video_378__q0000", "video_id": "video_378", "qid": 0, "question_type": "spatial understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_378_t5s.mp4"], "extra_info": {"question": "Which formation do the cheerleaders switch to when the music changes for the kick gesture?"}, "reward_model_ground_truth": {"multi_choice": ["A. They move from their oval formation to a straight line, quickly converging toward the center", "B. They maintain their oval formation but collectively take a step backward", "C. They move from two parallel straight lines to a single line by intersecting their positions", "D. They remain stationary, simply raising their pom-poms above their heads"], "correct_option": "A"}, "answer_text": "They move from their oval formation to a straight line, quickly converging toward the center", "reasoning_steps": [{"modality": "audio", "evidence": "The music changes for the kick gesture at 3:30.", "inference": "Cheerleaders begin the kick gesture at this time."}, {"modality": "vision", "evidence": "The cheerleaders are in oval formation at 3:23.", "inference": "They are in oval formation before the music changes."}, {"modality": "vision", "evidence": "The cheerleaders are converging toward the center at 3:30.", "inference": "They converge toward the center after the music changes."}, {"modality": "vision", "evidence": "The cheerleaders are in straight line at 3:34", "inference": "They move to a straight line after the music changes."}]}
{"id": "video_379__q0000", "video_id": "video_379", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_379_t5s.mp4"], "extra_info": {"question": "What is the fifth page of Nan's ppt like?"}, "reward_model_ground_truth": {"multi_choice": ["A. There are 2 clocks.", "B. There is a clock.", "C. There is a boy with a speaker in his eye.", "D. There is a picture of knives."], "correct_option": "A"}, "answer_text": "There are 2 clocks.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:34, he Black man said 'Let's see everyone's first slides.'", "inference": "Check everyone's first slide"}, {"modality": "audio", "evidence": "At 0:56, the Black man said 'Henriette and Nan' .", "inference": "first page of the Nan's PPT"}, {"modality": "audio", "evidence": "At 1:43, another man said 'Why don't we just focus on your presentation and we can see where we can help.'", "inference": "Next few slides will be Nan's slides."}, {"modality": "vision", "evidence": "At 2:45,  an image of two clocks and text of 'pie chart' appear.", "inference": "The fifth page of Nan's ppt is made with 2 clocks."}]}
{"id": "video_38__q0000", "video_id": "video_38", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_38_t5s.mp4"], "extra_info": {"question": "What will the actress do if she did not see the camera man she speaks to?"}, "reward_model_ground_truth": {"multi_choice": ["A. Finish the scene quickly.", "B. Act one more time.", "C. Ask why the director is so mean.", "D. Get angry and leave."], "correct_option": "A"}, "answer_text": "Finish the scene quickly.", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:48, the actress said 'Can you please leave?'", "inference": "The actress tried to get the camera man to leave."}, {"modality": "vision", "evidence": "At 1:50, the camera man shows up", "inference": "Pay attention to what she said to the cameraman."}, {"modality": "audio", "evidence": "At 1:53, the actress said '...in my eye line and I don't like that'.", "inference": "She dislikes having someone block her view. So if there were no photographers, she would finish the shot quickly."}]}
{"id": "video_38__q0001", "video_id": "video_38", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_38_t5s.mp4"], "extra_info": {"question": "What emotions did the director have on the actress finally?"}, "reward_model_ground_truth": {"multi_choice": ["A. Thankful", "B. Dislike", "C. Angry", "D. Indifferent"], "correct_option": "B"}, "answer_text": "Dislike", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:06, a man with a hat is shown in the scene.", "inference": "This is the director, identify him."}, {"modality": "vision", "evidence": "At 5:05, the man appeared with a a numb expression.", "inference": "He may be dissatisfied with the actress's performance."}, {"modality": "audio", "evidence": "At 5:08, the actress said 'Oh, you are all liked it that much.'.", "inference": "The actress is very satisfied with her own performance."}, {"modality": "audio", "evidence": "At 5:10, a woman answered 'No, we are trying to...'.", "inference": "Actually, the director dislike the actress's performance."}]}
{"id": "video_380__q0000", "video_id": "video_380", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_380_t5s.mp4"], "extra_info": {"question": "What is the source of conflict between Bob and his employees?"}, "reward_model_ground_truth": {"multi_choice": ["A. A disagreement over advertising ethics.", "B. The employees' desire to leverage the name for marketing vs Bob's deep shame.", "C. A simple misunderstanding of the name's negative connotations by the employees.", "D. Bob's inability to control his staff's radical ideas."], "correct_option": "B"}, "answer_text": "The employees' desire to leverage the name for marketing vs Bob's deep shame.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:17, his female employee says 'You gotta lean into the ISIS thing. It's good marketing tool.'", "inference": "She considers the brand as a marketing tool."}, {"modality": "audio", "evidence": "At 0:21, his male employee says 'I agree with Barbie.'", "inference": "He also agrees that the brand is a great resource to exploit."}, {"modality": "audio", "evidence": "At 0:26, Bob says 'No. We're not going to do this.'", "inference": "He disagrees with their opinion."}, {"modality": "vision", "evidence": "At 1:08, the car is written with 'terrorist'.", "inference": "It shows the marketing tool of relating the brand ISIS to the terrorist."}, {"modality": "audio", "evidence": "At 1:08, Bob says 'Stuff like that? Tha's all the crap.'", "inference": "Bob wants to distance his business from the terrorist group and regards it a shame."}]}
{"id": "video_380__q0001", "video_id": "video_380", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_380_t5s.mp4"], "extra_info": {"question": "What is the primary catalyst for Bob's escalating distress?"}, "reward_model_ground_truth": {"multi_choice": ["A. The financial cost of the vandalism.", "B. The employees' failure to grasp his shame.", "C. The problematic name of his employee, O.J. Anderson.", "D. A general lack of respect from the local community."], "correct_option": "B"}, "answer_text": "The employees' failure to grasp his shame.", "reasoning_steps": [{"modality": "vision", "evidence": "The video shows the text 'Bob Isis Owner, ISIS Toyota' at 0:03 of the video.", "inference": "Bob is the boss of ISIS Toyota."}, {"modality": "audio", "evidence": "One employee wants Bob to say ' we will be head the competition'.", "inference": "The employee is giving Bob suggestion."}, {"modality": "audio", "evidence": "Bob says'No, we're not doing this, what are you guys talking about'.", "inference": "Bob refuses to say those words."}, {"modality": "audio", "evidence": "The employee says he and another employee can hold AK-47.", "inference": "The employee keeps giving suggestion."}, {"modality": "audio", "evidence": "Bob says'no,no,no'.", "inference": "Bob also cannot accept this suggestion."}, {"modality": "audio", "evidence": "Another male emplyee askes Bob ' Can I be in a commercial?'.", "inference": "They are filming commercials."}, {"modality": "vision", "evidence": "At 1:39, the female employee takes out a a suicide bomb vest.", "inference": "They want Bob to wear it in the commercial."}, {"modality": "audio", "evidence": "Bob says 'I'm not wearing that'.", "inference": "Bob refused the suggestion again."}, {"modality": "audio", "evidence": "The employee says' why not'.", "inference": "The employee  ignores the Bob's refusal."}, {"modality": "vision", "evidence": "Bob pounces on the male employee in anger.", "inference": "His employees suggest ideas  pushe him from stressed to unhinged."}]}
{"id": "video_381__q0000", "video_id": "video_381", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_381_t5s.mp4"], "extra_info": {"question": "When Derwin says 'It's contagious energy',which player is to the left of the player?"}, "reward_model_ground_truth": {"multi_choice": ["A Player #3", "B Player #26", "C Player #40", "D Player #41"], "correct_option": "B"}, "answer_text": "Player #26", "reasoning_steps": [{"modality": "aduio", "evidece": "the question asks which player is to the left of Player #3 when he says 'It's contagious energy.'"}, {"modality": "vision", "evidece": "the video shows a group of Chargers players standing together. Player #3 is in the center. To his left is a player wearing #26."}]}
{"id": "video_382__q0000", "video_id": "video_382", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_382_t5s.mp4"], "extra_info": {"question": "How many people bagged a man?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "A"}, "answer_text": "1", "reasoning_steps": [{"modality": "audio", "evidence": "A man said 'You bag this fool, bag!'.", "inference": "After this point, a man will be bagged."}, {"modality": "vision", "evidence": "One man bagged a man at the 0:14 of the video.", "inference": "Only one person bagged a man."}]}
{"id": "video_382__q0001", "video_id": "video_382", "qid": 1, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_382_t5s.mp4"], "extra_info": {"question": "how many times did al gore's image appear?"}, "reward_model_ground_truth": {"multi_choice": ["A. 3", "B. 4", "C. 5", "D. 6"], "correct_option": "C"}, "answer_text": "5", "reasoning_steps": [{"modality": "audio", "evidence": "Singers mention 'Al Gore' in 1:19", "inference": "narrow the range."}, {"modality": "vision", "evidence": "The image of a middle-age man in suit titled Al Gore appeared.", "inference": "The man is Al Gore."}, {"modality": "vision", "evidence": "The image of Al Gore shows up in a blue background photo at timestamp 1:19.", "inference": "Al Gore's image appear for the first time."}, {"modality": "vision", "evidence": "The image of Al Gore shows up in a blue and red background photo making a gesture at timestamp 1:19.", "inference": "Al Gore's image appear for the second time."}, {"modality": "vision", "evidence": "The image of Al Gore shows up in a photo in blue shirt and tie at timestamp 1:20.", "inference": "Al Gore's image appear for the third time."}, {"modality": "vision", "evidence": "The image of Al Gore's full face photo shows up at timestamp 1:20.", "inference": "Al Gore's image appear for the forth time."}, {"modality": "vision", "evidence": "The image of Al Gore' s face appears in a tree at timestamp 1:31.", "inference": "Al Gore's image appear for the fifth time."}]}
{"id": "video_384__q0000", "video_id": "video_384", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_384_t5s.mp4"], "extra_info": {"question": "How many different instruments does the main singer play in total during the performance?"}, "reward_model_ground_truth": {"multi_choice": ["A: 1", "B: 2", "C: 3", "D: 4"], "correct_option": "C"}, "answer_text": "3", "reasoning_steps": [{"modality": "audio", "evidence": "Locate the main singers among the musicians", "inference": "Identifies who to focus on"}, {"modality": "vision", "evidence": "The first main singer is seen and heard switching from one instrument to another", "inference": "Confirms the first singer plays two different instruments"}, {"modality": "vision", "evidence": "The second main singer is seen and heard playing the same instrument throughout", "inference": "Confirms the second singer only plays one instrument"}, {"modality": "vision", "evidence": "Combine the counts from both singers (two from the first, one from the second)", "inference": "Total of three different instruments played"}]}
{"id": "video_386__q0000", "video_id": "video_386", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_386_t5s.mp4"], "extra_info": {"question": "For how many more beats does the music keep playing after the dancers stop, not counting the specific beat they stopped on?"}, "reward_model_ground_truth": {"multi_choice": ["A. 4", "B. 3", "C. 7", "D. 0"], "correct_option": "B"}, "answer_text": "3", "reasoning_steps": [{"modality": "vision", "evidence": "At timestamp 0:40, the dancers stop dancing on a downbeat.", "inference": "The beats need to be counted from then."}, {"modality": "audio", "evidence": "There are 4 beats including the downbeat.", "inference": "The likely answer is therefore 4."}, {"modality": "vision", "evidence": "At timestamp 0:40, this is the last 4/4 measure that is played.", "inference": "Not counting the downbeat, there are 3 beats after the dancing stops."}]}
{"id": "video_387__q0000", "video_id": "video_387", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_387_t5s.mp4"], "extra_info": {"question": "When the singing becomes harmonious, how many people are dancing?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 5"], "correct_option": "B"}, "answer_text": "3", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:09 of the video, the singing gradually becomes harmonious.", "inference": "The singing reaches a harmonious state at this timestamp."}, {"modality": "vision", "evidence": "At 0:09 of the video, there are three people dancing.", "inference": "Three people are dancing when the singing becomes harmonious."}]}
{"id": "video_388__q0000", "video_id": "video_388", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_388_t5s.mp4"], "extra_info": {"question": "When the proposal occurs, how many people are at the scene in addition to the couple getting engaged?"}, "reward_model_ground_truth": {"multi_choice": ["A 4", "B 5", "C 6", "D 7"], "correct_option": "D"}, "answer_text": "7", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:48, the man says, 'Do you accept my marriage proposal?'.", "inference": "It is a marriage proposal scene."}, {"modality": "vision", "evidence": "At 0:59, it shows 5 people applausing.", "inference": "5 people are at the scene."}, {"modality": "vision", "evidence": "At 1:01, it shows 2 skeletons applausing.", "inference": "2 skeletons are also at the scene,so there are 7 people."}]}
{"id": "video_389__q0000", "video_id": "video_389", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_389_t5s.mp4"], "extra_info": {"question": "How many people are wearing colors that are between the clothing colors of the shop clerk he passed when leaving Herb a message?"}, "reward_model_ground_truth": {"multi_choice": ["A 4", "B 5", "C 6", "D 7"], "correct_option": "C"}, "answer_text": "6", "reasoning_steps": [{"modality": "audio", "evidence": "When leaving Herb a message, he passes a shop .", "inference": "Mark the location through the message."}, {"modality": "vision", "evidence": "The clerk at the shop he passed when leaving Herb a message was wearing clothing with two colors: red and green.", "inference": "People should wear in red or green."}, {"modality": "vision", "evidence": "At 0:01, a woman wears red swimsuit.", "inference": "There are 1 person wearing in  in red or green so far."}, {"modality": "vision", "evidence": "At 0:02, a person wears hat also wears a red underwear.", "inference": "There are 2 people wearing in  in red or green so far."}, {"modality": "vision", "evidence": "At 0:03, a black-hair woman wears a red underwear.", "inference": "There are 3 people wearing  in red or green so far."}, {"modality": "vision", "evidence": "At 0:06, a woman wears a red bra.", "inference": "There are 4 people wearing in red or green so far."}, {"modality": "vision", "evidence": "At 0:06, a man wears a red belt.", "inference": "There are 5 people wearing in red or green so far."}, {"modality": "vision", "evidence": "At 0:10, a woman wears red bra.", "inference": "There are 6 people wearing  in red or green so far."}]}
{"id": "video_393__q0000", "video_id": "video_393", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_393_t5s.mp4"], "extra_info": {"question": "In the first demonstration for makam Segah, which of the played notes is played with vibrato?"}, "reward_model_ground_truth": {"multi_choice": ["A. E", "B. F", "C. G", "D. B"], "correct_option": "D"}, "answer_text": "B", "reasoning_steps": [{"modality": "vision", "evidence": "the demonstrations are announced at timestamp 2:40.", "inference": "the demonstration for makam Segah likely follows."}, {"modality": "vision", "evidence": "Segah is visible next to the video of the keyboard starting at timestamp 2:48.", "inference": "this is the demonstration for makam Segah."}, {"modality": "vision", "evidence": "the score displays 5 notes in the makam Segah.", "inference": "one of these is likely the one being played with vibrato."}, {"modality": "audio", "evidence": "the last note in makam Segah is played with vibrato.", "inference": "this is the note in question."}, {"modality": "audio", "evidence": "the last note played in makam Segah is B.", "inference": "the answer is B."}]}
{"id": "video_394__q0000", "video_id": "video_394", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_394_t5s.mp4"], "extra_info": {"question": "Why does the boss cancel the decision to fire this employee?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because firing this employee is too much trouble.", "B. Because this employee still has value to the company.", "C. Because this employee intends to sue the company.", "D. Because this boss cannot afford to pay severance pay."], "correct_option": "A"}, "answer_text": "Because firing this employee is too much trouble.", "reasoning_steps": [{"modality": "vision", "evidence": "From vision, the crying person says 'I am fired?'", "inference": "The crying person is the one who got fired, and the person using the computer is the boss."}, {"modality": "audio", "evidence": "Start from 0:11 of the video,the employer begins to explain the complex, legally mandated procedures for termination in France.", "inference": "In France, the process of firing an employee is quite complicated."}, {"modality": "audio", "evidence": "Start from 0:50 of the video,the employer stated that the bureaucratic process is more of a 'pain in the ass' than simply keeping the employee on staff.", "inference": "The hassle outweighs the benefit of dismissal."}]}
{"id": "video_394__q0001", "video_id": "video_394", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_394_t5s.mp4"], "extra_info": {"question": "When the boss said 'you are fired' for the last time, which country was the label under the table from?"}, "reward_model_ground_truth": {"multi_choice": ["A. America", "B. Australia.", "C. France.", "D. English."], "correct_option": "A"}, "answer_text": "America", "reasoning_steps": [{"modality": "audio", "evidence": "The boss said, 'you are fired.'for the last time at 1:04.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "The letter under the table is 'America'.", "inference": "The letter under the table is 'America'."}]}
{"id": "video_395__q0000", "video_id": "video_395", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_395_t5s.mp4"], "extra_info": {"question": "Which statement accurately compares the types of produce and their respective sources in the woman's harvest basket?"}, "reward_model_ground_truth": {"multi_choice": ["A. The majority of the vegetables like lettuce and carrots were harvested from her main garden, while fruits such as apples were collected from a separate orchard.", "B. Artichokes, apples, carrots, lettuce, and garlic were cultivated in her primary garden, but the cucumbers were sourced from her neighbor Nick's distinct garden patch.", "C. All garden produce, including the artichokes, carrots, and cucumbers, originated from her personal raised beds, and the eggs were from her chickens.", "D. She exclusively harvested fruits like apples and vegetables like lettuce from her own garden, acquiring specialized produce such as garlic and cucumbers from a nearby farm."], "correct_option": "B"}, "answer_text": "Artichokes, apples, carrots, lettuce, and garlic were cultivated in her primary garden, but the cucumbers were sourced from her neighbor Nick's distinct garden patch.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:10, a woman's voice says' Firstly, I am going to grab some artichokes.'", "inference": "The woman harvests artichokes."}, {"modality": "audio", "evidence": "A 0:14, a woman's voice says ' Next, I am going to grab some apples ready to harvest.'", "inference": "The woman harvests some apples."}, {"modality": "vision", "evidence": "At 0: 20, a hand puts some green apples in the basket.", "inference": "The person graps the apple."}, {"modality": "vision", "evidence": "At 0: 26, the person holds a carrot in hand.", "inference": "The person harvests a carrot."}, {"modality": "vision", "evidence": "At 0:37, the woman holds many carrots in hand.", "inference": "The woman harvests some carrots."}, {"modality": "audio", "evidence": "At 0: 57, a woman's voice says' I think I am going to  harvest some of these gorgeous red lettuces.'", "inference": "The woman will harvest some lettuces."}, {"modality": "audio", "evidence": "At 1: 07, a woman's voice says ' OK, let's snag over into next garden to grab some cucumbers.'", "inference": "The woman will go to other's garfen to get cucumbers."}, {"modality": "audio", "evidence": "At 1:10, the woman says'This little patch is his garden.'", "inference": "The woman go to other's garden."}, {"modality": "vision", "evidence": "At 1:31, some garlics are shown in the footage.", "inference": "The woman harvests some garlics."}]}
{"id": "video_396__q0000", "video_id": "video_396", "qid": 0, "question_type": "temporal understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_396_t5s.mp4"], "extra_info": {"question": "After the male singer starts singing, what actions do the dancers take?"}, "reward_model_ground_truth": {"multi_choice": ["A: They repeat the same movement 12 times without change", "B: They repeat one movement 4 times, then repeat it 9 more times at a different speed", "C: They repeat one movement 4 times, then switch to a completely new movement repeated 9 times", "D: They repeat one movement 5 times, then repeat it 8 more times at a different speed"], "correct_option": "B"}, "answer_text": "They repeat one movement 4 times, then repeat it 9 more times at a different speed", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:01, the male singer's voice is first heard.", "inference": "Marks the moment when dancer actions should be checked."}, {"modality": "vision", "evidence": "At 1:02, the dancers perform one movement and repeat it 4 times.", "inference": "Identifies the first set of repetitions."}, {"modality": "vision", "evidence": "At 1:40, the dancers continue the same movement but faster, repeating it 9 more times.", "inference": "Identifies the second set of repetitions at a different speed."}]}
{"id": "video_397__q0000", "video_id": "video_397", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_397_t5s.mp4"], "extra_info": {"question": "What best describes the third party to the desk?"}, "reward_model_ground_truth": {"multi_choice": ["A. A young couple asking if they can borrow a scooter overnight.", "B. Two young women asking for a spare key to their friend's room.", "C. A young man complaining that he did not get the Copa Suite he paid for.", "D. A young woman asking for directions to find a psychic or a fortune teller."], "correct_option": "A"}, "answer_text": "A young couple asking if they can borrow a scooter overnight.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:10, a woman wearing a pink headscarf appears in front of the desk.", "inference": "This is the first party to the desk."}, {"modality": "vision", "evidence": "At 0:35, a man with sunglasses appears in front of the desk.", "inference": "This is the secong party to the desk."}, {"modality": "vision", "evidence": "At 1:20, the young cople shows in front of the desk.", "inference": "They are the third party to the desk."}, {"modality": "audio", "evidence": "The female of the couple says'Borrow a Vespa scooter over the night.'.", "inference": "So the answer is 'A young couple asking if they can borrow a scooter overnight.'."}]}
{"id": "video_397__q0001", "video_id": "video_397", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_397_t5s.mp4"], "extra_info": {"question": "Why the guest is denied the 'Copa Suite' he paid for?"}, "reward_model_ground_truth": {"multi_choice": ["A. The receptionist believes the guest cannot afford the suite, similar to the first guest who had no credit card.", "B. The hotel has likely overbooked the room, a problem consistent with the staff's inability to manage guest requests.", "C. The staff is intentionally withholding the room out of spite for the guest's entitled attitude from the moment he arrived.", "D. The suite is unavailable because it is part of an active police investigation involving a body found on the beach."], "correct_option": "B"}, "answer_text": "The hotel has likely overbooked the room, a problem consistent with the staff's inability to manage guest requests.", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:57, there are constant guests coming to check in at the front desk.", "inference": "The hotel may receive overbooked guests."}, {"modality": "vision", "evidence": "The hotel environment and staff are chaotic and mismanaged.", "inference": "The staff may unable to manage guest requests."}, {"modality": "audio", "evidence": "At 3:01, the guest askes why that so damn hard to get the 'Copa Suite'.", "inference": "The guest is denied the 'Copa Suite' he paid for."}, {"modality": "audio", "evidence": "The receptionist tells the guest that the suite 'is just not available'.", "inference": "This may be caused by a logistical problem."}]}
{"id": "video_398__q0000", "video_id": "video_398", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_398_t5s.mp4"], "extra_info": {"question": "How many jokes does Jim tries before realizing that meatballs work on Stanley and does Stanley really hate meatballs?"}, "reward_model_ground_truth": {"multi_choice": ["A 2, yes", "B 2, no", "C 3, yes", "D 3, no"], "correct_option": "D"}, "answer_text": "3, no", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:33, Stanley puts down his briefcase and the table sinks down.", "inference": "It is the first joke."}, {"modality": "audio", "evidence": "At 0:49, Stanley says'Jim, come on. So juvenile.'.", "inference": "It is the second joke."}, {"modality": "audio", "evidence": "At 1:13, Stanley says'Very funny.'.", "inference": "It is the third joke."}, {"modality": "audio", "evidence": "At 1:35, Stanley says 'That idiot's been feeding us for a week.'.", "inference": "It dedicates that Stanley does not really hate meatballs."}]}
{"id": "video_399__q0000", "video_id": "video_399", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_399_t5s.mp4"], "extra_info": {"question": "Who is Willem talking to after turning around?"}, "reward_model_ground_truth": {"multi_choice": ["A. A male passerby.", "B. The vlogger.", "C. A female passerby.", "D. A colleague."], "correct_option": "C"}, "answer_text": "A female passerby.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:05, a man said his name is 'Willem'.", "inference": "He is the man mentioned in the question."}, {"modality": "vision", "evidence": "At 0:05, Willem is in the scene.", "inference": "Identify this man."}, {"modality": "vision", "evidence": "At 0:28, his head jerks to the right, away from the interviewer.", "inference": "He is not speaking to the person filming the video."}, {"modality": "audio,", "evidence": "At 00:31, Willem suddenly erupts in loud, genuine laughter.", "inference": "He may talks to another people."}, {"modality": "audio", "evidence": "A female voice can be heard at 0:35.", "inference": "Willem is talking to her, a female passerby."}]}
{"id": "video_399__q0001", "video_id": "video_399", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_399_t5s.mp4"], "extra_info": {"question": "How does Willem Dafoe generally position himself and move in relation to the camera during the interaction?"}, "reward_model_ground_truth": {"multi_choice": ["A. He primarily walks forward, moving closer to the camera.", "B. He remains mostly stationary, facing the camera directly, throughout the entire video.", "C. He walks away from the camera's general direction while frequently turning his head to maintain eye contact.", "D. He consistently walks backward, maintaining a fixed distance from the camera."], "correct_option": "C"}, "answer_text": "He walks away from the camera's general direction while frequently turning his head to maintain eye contact.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:05, the man says his name is Willem.", "inference": "Identify who is Willem, the person mentioned in the question."}, {"modality": "vision", "evidence": "At 0:05, Willem stops.", "inference": "Before this moment, Willem keeps walking away from the camera's general direction."}, {"modality": "vision", "evidence": "At 0:05, Willem looks at the camera.", "inference": "He maintains eye contact."}, {"modality": "vision", "evidence": "At 0:12, Willem moves forward away the camera's general direction.", "inference": "This is his first movement."}, {"modality": "vision", "evidence": "At 0:12, Willem looks at the camera again.", "inference": "He maintains eye contact again."}, {"modality": "vision", "evidence": "At 0:18, Willem moves forward away the camera's general direction.", "inference": "This is his second movement."}, {"modality": "vision", "evidence": "At 0:19, Willem looks at the camera again.", "inference": "He maintains eye contact again."}, {"modality": "vision", "evidence": "At 0:32, Willem moves forward away the camera's general direction.", "inference": "This is his third movement."}, {"modality": "vision", "evidence": "At 0:34, Willem looks at the camera again.", "inference": "He maintains eye contact again."}]}
{"id": "video_4__q0000", "video_id": "video_4", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_4_t5s.mp4"], "extra_info": {"question": "In which position next to Zack is the individual who identifies Zack's aunt located?"}, "reward_model_ground_truth": {"multi_choice": ["A.the 1st", "B.the 2nd", "C.the 3rd", "D.the 4th"], "correct_option": "C"}, "answer_text": "the 3rd", "reasoning_steps": [{"modality": "audio", "evidence": "At the beginning of the show, the host introduces the contestants.", "inference": "We can hear Zack being introduced by name."}, {"modality": "vision", "evidence": "identify zack.", "inference": "pay attention to him."}, {"modality": "audio", "evidence": "the host introduces another contestant and explicitly states that she is Zack's aunt, Angela.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "identify zack's aunt.", "inference": "pay attention to her."}, {"modality": "vision", "evidence": "count the people standing to the left of Zack.", "inference": "we can determine that the woman previously identified as his aunt is in the third position to his left."}]}
{"id": "video_40__q0000", "video_id": "video_40", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_40_t5s.mp4"], "extra_info": {"question": "When someone said 'let me guess,' what was the first item on the left side of the table on the left side of the screen??"}, "reward_model_ground_truth": {"multi_choice": ["A.Green jar.", "B.Green vase.", "C.Table lamp.", "D.Clock."], "correct_option": "C"}, "answer_text": "Table lamp.", "reasoning_steps": [{"modality": "audio", "evidence": "At 01:10, Spider-Man said 'let me guess.'", "inference": "Identify the sentence."}, {"modality": "vision", "evidence": "At 01:10, there is a table on the left side of the screen.", "inference": "Identify the table on the left side of the screen."}, {"modality": "vision", "evidence": "At 01:10, the first item on the left side of the table is a table lamp.", "inference": "Identify the first item on the left side of the table. So it's a table lamp."}]}
{"id": "video_40__q0001", "video_id": "video_40", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_40_t5s.mp4"], "extra_info": {"question": "When someone said 'let me guess', which of the following items was on the table on the left side of the screen?"}, "reward_model_ground_truth": {"multi_choice": ["A.Book.", "B.Green vase.", "C.Table lamp.", "D.Red jar."], "correct_option": "C"}, "answer_text": "Table lamp.", "reasoning_steps": [{"modality": "audio", "evidence": "At 01:10, Spider-Man said 'let me guess.'", "inference": "Identify the sentence."}, {"modality": "vision", "evidence": "At 01:10, there is a table on the left side of the screen.", "inference": "Identify the table on the left side of the screen."}, {"modality": "vision", "evidence": "At 01:10, the items on the table is a table lamp, a square decoration and a triangular decoration.", "inference": "Identify one of the items on the table. So it's a table lamp."}]}
{"id": "video_400__q0000", "video_id": "video_400", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_400_t5s.mp4"], "extra_info": {"question": "When the girl says 'you can't think of anyone weirder?', is the couple she is referring to the same as the couple that the girl in red is referring to, and what does Leonard think of the girl's paper?"}, "reward_model_ground_truth": {"multi_choice": ["A yes, good", "B no, good", "C yes, bad", "D no, bad"], "correct_option": "D"}, "answer_text": "no, bad", "reasoning_steps": [{"modality": "audio", "evidence": "When the girl says 'you can't think of anyone weirder?'.", "inference": "The girl wants to know if there is someone any weirder."}, {"modality": "vision", "evidence": "When the girl in red answers this question, she looks at the girl wearing green.", "inference": "The answer is no."}, {"modality": "audio", "evidence": "Leonard  says 'she writes like she cooks'.", "inference": "The answer can be found from Penny's cooking."}, {"modality": "audio", "evidence": "The water was not boiling when Penny cooked the spaghetti.", "inference": "The answer is bad."}]}
{"id": "video_401__q0000", "video_id": "video_401", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_401_t5s.mp4"], "extra_info": {"question": "When the music in the video changes from urgent to soothing, what instrument appears?"}, "reward_model_ground_truth": {"multi_choice": ["A: Piano", "B: Drum", "C: Qanun", "D: Santur"], "correct_option": "B"}, "answer_text": "Drum", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:53, the music changes from urgent to soothing.", "inference": "The music tempo and mood transition occurs at this timestamp."}, {"modality": "vision", "evidence": "At 0:53, the first instrument that appears is the drum.", "inference": "The drum is the primary instrument visible during the music transition."}]}
{"id": "video_403__q0000", "video_id": "video_403", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_403_t5s.mp4"], "extra_info": {"question": "When someone suggesting to take a better photo by 'bouncing,' what pose was Jimmy making?"}, "reward_model_ground_truth": {"multi_choice": ["A. He was standing next to Zendaya.", "B. He was crouching down behind Zendaya.", "C. He was gesturing towards the camera.", "D. He was handing microphone to the camera."], "correct_option": "B"}, "answer_text": "He was crouching down behind Zendaya.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:19, Jimmy says 'she can bounce me'.", "inference": "They will make actions for filming."}, {"modality": "audio", "evidence": "At 0:24, people holding the phone says 'down up down up'.", "inference": "He is guiding Zendaya to make gestures."}, {"modality": "vision", "evidence": "At 0:25, Zendaya moves her hands up and down.", "inference": "Zendaya starts to 'bounce' Jimmy."}, {"modality": "vision", "evidence": "At 0:28, Jimmy starts to crouch down behind Zendays.", "inference": "It shows Jimmy's specific posture of crounching down."}]}
{"id": "video_404__q0000", "video_id": "video_404", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_404_t5s.mp4"], "extra_info": {"question": "What action does the singer continue doing while getting out of bed to dance?"}, "reward_model_ground_truth": {"multi_choice": ["A. Singing", "B. Swinging hands", "C. Finger snapping", "D. Playing an instrument"], "correct_option": "C"}, "answer_text": "Finger snapping", "reasoning_steps": [{"modality": "vision", "evidence": "The singer visibly sings and smokes while in bed.", "inference": "Likely continues one of these actions."}, {"modality": "audio", "evidence": "There are finger snaps while the singer is singing, which are not well-synchronized with the instrumental track.", "inference": "The singer is likely performing the finger snaps."}, {"modality": "vision", "evidence": "The singer has a cigarette in her mouth while dancing.", "inference": "Likely not singing or sighing while dancing."}, {"modality": "audio", "evidence": "The finger snaps continue even when the singer isn't singing.", "inference": "Confirms the singer continues finger snapping even while getting out of bed to dance."}]}
{"id": "video_405__q0000", "video_id": "video_405", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_405_t5s.mp4"], "extra_info": {"question": "What is the primary intention of the animal keeper in presenting the information and animal interactions shown in this video?"}, "reward_model_ground_truth": {"multi_choice": ["A. To educate viewers about the natural digging behaviors of various animals and how the park supports them.", "B. To demonstrate the daily feeding routines and the types of food provided to the animals at the park.", "C. To showcase the variety of animals housed at the Wildlife Prairie Park and encourage visits.", "D. To highlight the unique characteristics and general habits of red foxes, badgers, and raccoons."], "correct_option": "A"}, "answer_text": "To educate viewers about the natural digging behaviors of various animals and how the park supports them.", "reasoning_steps": [{"modality": "vision", "evidence": "A woman shows up in the scene  with the subtitles ' Lead Animal Keeper' at 0:03.", "inference": "The woman is the animal keeper."}, {"modality": "audio", "evidence": "The woman says 'today we will talk about digging' at 0:08", "inference": "The woman will talk about animal digging next."}, {"modality": "vision", "evidence": "The subtitles 'DIGGING' shows up on the scene.", "inference": "The animal keeper intent to educate viewers about the natural digging behaviors of various animals."}, {"modality": "audio", "evidence": "The woman says she is hiding the food of raccoon at 1: 21.", "inference": "She hides the food instead of raccoon's digging."}, {"modality": "audio", "evidence": "The says this will help exhibit raccoon's digging.", "inference": "She is demonstrating how the park supports the animals."}]}
{"id": "video_405__q0001", "video_id": "video_405", "qid": 1, "question_type": "summarization", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_405_t5s.mp4"], "extra_info": {"question": "During which of the keeper's interactions with animals was the background music used instead of her spoken commentary?"}, "reward_model_ground_truth": {"multi_choice": ["A. During the interaction with the Ember.", "B. During the interaction with the Mr. B.", "C. During the interaction with the Ruby.", "D. During the interaction with the Badger."], "correct_option": "C"}, "answer_text": "During the interaction with the Ruby.", "reasoning_steps": [{"modality": "vision", "evidence": "At 1:50, the keeper feeds Raccoon (Ruby).", "inference": "Point out the key period that needs attention."}, {"modality": "vision", "evidence": "At 1:52,  the raccoon begins to explore the bowl of food.", "inference": "Pay attention to the background music."}, {"modality": "audio", "evidence": "At 1:52, the keeper's verbal commentary stops completely.", "inference": "While interacting with Ruby, keeper doesn't say anything."}, {"modality": "audio", "evidence": "At 1:52, there's only the background music.", "inference": "The background music is used instead of keeper's spoken commentary."}]}
{"id": "video_407__q0000", "video_id": "video_407", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_407_t5s.mp4"], "extra_info": {"question": "Who killed the roach that said 'Give it a minite'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Bill", "B. Bill's son", "C. Bill's wife", "D. no one"], "correct_option": "D"}, "answer_text": "no one", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:03, the man said 'There's nothing worse than unexpected house guests, especially roaches.'", "inference": "The guests will be roaches."}, {"modality": "vision", "evidence": "At 0:08, the black person with a cockroach-like appearance is shown.", "inference": "Find the black roach."}, {"modality": "vision", "evidence": "At 0:11, the white person with a cockroach-like appearance is shown.", "inference": "Find the white roach."}, {"modality": "audio", "evidence": "At 0:18, the white roach said: 'Give it a minite'.", "inference": "Find the roach metioned in the question."}, {"modality": "vision", "evidence": "At 2:15, the black roach that is killed.", "inference": "No one killed the white roach that said 'Give it a minite'."}]}
{"id": "video_408__q0000", "video_id": "video_408", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_408_t5s.mp4"], "extra_info": {"question": "What instrument is the musician sitting next to the main singer playing?"}, "reward_model_ground_truth": {"multi_choice": ["A: Santur", "B: Qanun", "C: Setar", "D: Tombak"], "correct_option": "A"}, "answer_text": "Santur", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:00, a woman is singing with a microphone.", "inference": "The woman is the main singer."}, {"modality": "vision", "evidence": "At 1:08, the musicians who are playing Santur seats next to her.", "inference": "The instrument is Santur."}]}
{"id": "video_409__q0000", "video_id": "video_409", "qid": 0, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_409_t5s.mp4"], "extra_info": {"question": "Why does the background music suddenly become scary?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because it can not return to its position in time.", "B. Because it is afraid of being broken.", "C. Because it is discovered that it could move.", "D. Because it was afraid of being eaten by the horse."], "correct_option": "A"}, "answer_text": "Because it can not return to its position in time.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:43, the little chicken figurine looks at its watch.", "inference": "The little chicken figurine needs to pay attention to the time."}, {"modality": "vision", "evidence": "At 0:43, the watch shows the start and end time for work.", "inference": "The little chicken figurine should return to the original position on time."}, {"modality": "audio", "evidence": "At 0:45, the background music suddenly becomes scary.", "inference": "The music changes to a frightening tone at this specific timestamp."}, {"modality": "vision", "evidence": "At 0:45, the little chicken figurine imagines a lot of trouble happened if he doesn't in its original booth.", "inference": "The little chicken figurine is afraid that he can't return in time."}]}
{"id": "video_409__q0001", "video_id": "video_409", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_409_t5s.mp4"], "extra_info": {"question": "What best describes the primary motivation behind the museum artifacts' actions after the museum closes?"}, "reward_model_ground_truth": {"multi_choice": ["A. To recover valuable items that have been stolen from the museum.", "B. To avoid detection by museum staff while they enjoy their secret activities.", "C. To entertain visitors by coming alive and performing various acts during opening hours.", "D. To relive their past lives and historical roles outside the museum's confines."], "correct_option": "B"}, "answer_text": " To avoid detection by museum staff while they enjoy their secret activities.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:03 of the video, the words shown on the screen are 'After the museum artifacts get off the work'", "inference": "The clip shown after is about what will those museum artifacts do after the museum closes."}, {"modality": "audio", "evidence": "At 0:15 of the video, one of the museum artifacts says to another;'Go for fun'", "inference": "They will enjoy themselves after the museum closes."}, {"modality": "vision", "evidence": "From 0:32 of the video, the museum artifacts dance together.", "inference": "They enjoy themselves together."}, {"modality": "vision", "evidence": "At 0:42 of the video, one of the museum artifacts wakes up late.", "inference": "It played so engrossedly that it lost itself in it."}, {"modality": "vision", "evidence": "At 0:47 of the video, the screen shows a news report that the museum artifact has diappeared mysteriously.", "inference": "The detection by meseum staff doesn't know that the museum artifacts' actions after the meseum closes."}]}
{"id": "video_41__q0000", "video_id": "video_41", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_41_t5s.mp4"], "extra_info": {"question": "After the anchor said 'Scout is trying to now walk forward with Xun', whose kill was the third one that gala got?"}, "reward_model_ground_truth": {"multi_choice": ["A.Xun.", "B.Scout.", "C.Xiaoxu.", "D.Peyz."], "correct_option": "A"}, "answer_text": "Xun.", "reasoning_steps": [{"modality": "audio", "evidence": "At 05:28, the anchor said 'Scout is trying to now walk forward with Xun'.", "inference": "Pay attention to the third kill gala got."}, {"modality": "vision", "evidence": "At 05:50, gala got the third kill of Xun.", "inference": "Xun's kill was the third one that gala got."}]}
{"id": "video_41__q0001", "video_id": "video_41", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_41_t5s.mp4"], "extra_info": {"question": "When the anchor said 'now scout is compromised, already use the flash', who got the kill on the scout?"}, "reward_model_ground_truth": {"multi_choice": ["A.Gala.", "B.TheShy.", "C.Meiko.", "D.Rookie."], "correct_option": "A"}, "answer_text": "Gala.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:47, the anchor said 'now scout is compromised, already use the flash'.", "inference": "Pay attention to the one got the kill on the scout."}, {"modality": "vision", "evidence": "At 02:50, gala got the kill.", "inference": "When the anchor said 'now scout is compromised, already use the flash', gala got the kill on the scout."}]}
{"id": "video_41__q0002", "video_id": "video_41", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_41_t5s.mp4"], "extra_info": {"question": "After the anchor said 'Scout is trying to now walk forward with Xun', whose kill was the second one that gala got?"}, "reward_model_ground_truth": {"multi_choice": ["A.Xun.", "B.Scout.", "C.Xiaoxu.", "D.Peyz."], "correct_option": "C"}, "answer_text": "Xiaoxu.", "reasoning_steps": [{"modality": "audio", "evidence": "At 05:28, the anchor said :' Scout is trying to now walk forward with Xun'.", "inference": "Pay attention the second kill that gala got."}, {"modality": "vision", "evidence": "At 05:50, gala got the second kill of Xiaoxu.", "inference": "Xiaoxu's kill was the second one that gala got."}]}
{"id": "video_410__q0000", "video_id": "video_410", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_410_t5s.mp4"], "extra_info": {"question": "What animal did Spotty turn the punisher of the game into?"}, "reward_model_ground_truth": {"multi_choice": ["A From snakes to dinosaurs", "B From snakes to chickens", "C From dinosaurs to chickens", "D From dinosaurs to snakes"], "correct_option": "A"}, "answer_text": "From snakes to dinosaurs", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:20 in the video, the rabbit says he wants to turn the snakes into 'a dinosaur!'", "inference": "The snake will be turned into a dinosaur next accrording to the rabbit."}, {"modality": "vision,", "evidence": "At 1:26  in the video, the punisher of the game changes from a snake into a dinosaur after the rabbit speaks.", "inference": "The rabbit turned the punisher of the game from snakes to dinosaurs."}, {"modality": "audio", "evidence": "At 1:55 in the video,the boy calls the rabbit Spotty, which proves that the rabbit is Spotty.", "inference": "The name of this rabbit is Spotty."}]}
{"id": "video_411__q0000", "video_id": "video_411", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_411_t5s.mp4"], "extra_info": {"question": "Summer's escalating annoyance and eventual departure from the pool area are most directly caused by?"}, "reward_model_ground_truth": {"multi_choice": ["A. The Cogs repeatedly splashing her and eating all the eclair.", "B. Rick's disinterest in spending quality time with her and the Cogs' subservient, attention-seeking behavior.", "C. Her inability to join Rick and the Cogs in the pool due to the lack of space.", "D. Rick's constant consumption of alcoholic beverages and his disregard for her feelings."], "correct_option": "B"}, "answer_text": "Rick's disinterest in spending quality time with her and the Cogs' subservient, attention-seeking behavior.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:09, Summer has the an angry expression.", "inference": "She may be dissatisfied with some of Rick's behaviors."}, {"modality": "audio", "evidence": "At 0:50, Summer says 'I wanted to be at a restaurant with you.'.", "inference": "This indicates that Summer Summer originally wants to go to the restaurant with Rick for dinner."}, {"modality": "audio", "evidence": "At 1:15, Summer says 'I hate these Cogs asshole!'.", "inference": "It indicates that Summer dislike Cogs' subservient, attention-seeking behavior."}, {"modality": "vision", "evidence": "At 0:17, Summer leaves angrily.", "inference": "This shows that Rick is not interested in spending time with her."}]}
{"id": "video_411__q0001", "video_id": "video_411", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_411_t5s.mp4"], "extra_info": {"question": "How many people predicted Summer's response at the same time in the swimming pool?"}, "reward_model_ground_truth": {"multi_choice": ["A.1", "B.2", "C.3", "D.4"], "correct_option": "D"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "Rick asks summer to get the jell-o shots at 1:11.", "inference": "bring about the topic Summer will respond to."}, {"modality": "audio", "evidence": "At 1:16, After Summer said 'I', three robots said 'hate these cock assholes'.", "inference": "Three robots predicted Summer's reponse."}, {"modality": "audio", "evidence": "At 1:15, After Summer said 'I', Rick also said 'hate these cock assholes'.", "inference": "Rick also predicted Summer's response."}, {"modality": "vision", "evidence": "At 1:15, there are 4 people in the swimming pool.", "inference": "The swimming pool contains 4 individuals at this timestamp."}]}
{"id": "video_412__q0000", "video_id": "video_412", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_412_t5s.mp4"], "extra_info": {"question": "What did the food dude as a model of Cocoon say?"}, "reward_model_ground_truth": {"multi_choice": ["A. Hurry up with that food!", "B. The other 3 of us are starving.", "C. We are noticably overweight.", "D. Piping hot"], "correct_option": "C"}, "answer_text": "We are noticably overweight.", "reasoning_steps": [{"modality": "vision", "evidence": "At 1:36, the food dude with orange hair was used as a model of Cocoon.", "inference": "What he said should be around this moment."}, {"modality": "audio", "evidence": "At 1:27, the food dude with orange hair said 'We are noticably overweight.'.", "inference": "the food dude as a model of Cocoon said 'We are noticably overweight.'."}]}
{"id": "video_413__q0000", "video_id": "video_413", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_413_t5s.mp4"], "extra_info": {"question": "How did Aladdin feel like after he knew boyfriends?"}, "reward_model_ground_truth": {"multi_choice": ["A. happy", "B. angry", "C. indifferent", "D. insecure"], "correct_option": "D"}, "answer_text": " insecure", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:42, when the girl mentions she had many boyfriends before, the boy's facial expression becomes tense.", "inference": "The boy shows visible signs of nervousness and concern about the girl's dating history."}, {"modality": "audio", "evidence": "At 1:00, Aladdin expresses concern that he 'can't handle' her physically and worries about not being 'enough man.'", "inference": "Aladdin's words and tones appear to feel insecure, he feels insecure."}]}
{"id": "video_414__q0000", "video_id": "video_414", "qid": 0, "question_type": "summarization", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_414_t5s.mp4"], "extra_info": {"question": "Where did the jewelry end up?"}, "reward_model_ground_truth": {"multi_choice": ["A. Hidden by bandits.", "B. Found by the police.", "C. Hidden by Agnes.", "D. Stolen by the Minions."], "correct_option": "D"}, "answer_text": "Stolen by the Minions.", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:44 seconds of the video, the Minions are very proud.", "inference": "The Minions are very happy and proud."}, {"modality": "vision", "evidence": "At 3:44 seconds in the video, the Minion is carrying stolen jewelry.", "inference": "The jewelry ultimately fell into the hands of the Minions."}]}
{"id": "video_414__q0001", "video_id": "video_414", "qid": 1, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_414_t5s.mp4"], "extra_info": {"question": " After the little girl gets what she wanted, what happens to the other thing mentioned in the newspaper?"}, "reward_model_ground_truth": {"multi_choice": ["A. Hidden by bandits.", "B. Found by the police.", "C. Hidden by Agnes.", "D. Stolen by the Minions."], "correct_option": "D"}, "answer_text": "Stolen by the Minions.", "reasoning_steps": [{"modality": "vision", "evidence": "At 3:42, the newspaper shows 'JEWELRY STILL MISSING'.", "inference": "Though theives were caught, police still didn't find the jewelry."}, {"modality": "audio", "evidence": "The little girl has been wanting to eat ice cream.", "inference": "Ice cream is the item the little girl desires."}, {"modality": "vision", "evidence": "At 3:43, Minions wear those jewelry.", "inference": "Minions stole the jewelry."}]}
{"id": "video_415__q0000", "video_id": "video_415", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_415_t5s.mp4"], "extra_info": {"question": "When who used home test appeared for the fourth time, what did she say?"}, "reward_model_ground_truth": {"multi_choice": ["A. I had to isolate from my 3 kids.", "B. Vanks, Vovid.", "C. Watch the netflix I want to watch.", "D. Again?"], "correct_option": "A"}, "answer_text": "I had to isolate from my 3 kids.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:07 in this video, the woman  was first shown stressed out by her children.", "inference": "This is her first appearance."}, {"modality": "vision", "evidence": "At 0:18 in this video, the woman was happily dancing in her living room.", "inference": "This is her second appearance."}, {"modality": "vision", "evidence": "At 0:32 in this video,the woman was sitting on the couch.", "inference": "This is her third appearance."}, {"modality": "vision", "evidence": "At 1:03 in this video, the woman was waving hands to her 3 kids and husband.", "inference": "This is her fourth appearance."}, {"modality": "audio", "evidence": "She said, 'And of course, I had to isolate from my 3 kids.'.", "inference": "She said,' I had to isolate from my 3 kids.'"}]}
{"id": "video_416__q0000", "video_id": "video_416", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_416_t5s.mp4"], "extra_info": {"question": "Who is on the right-hand side of the person who needs a hotel that is not family-friendly?"}, "reward_model_ground_truth": {"multi_choice": ["A. The person wearing a pink hat.", "B. The person wearing a floral shirt.", "C. The person with loose curly hair.", "D. The person holding a blue phone."], "correct_option": "C"}, "answer_text": "The person with loose curly hair.", "reasoning_steps": [{"modality": "audio,", "evidence": "A lady expressed a preference for a hotel that is less family-friendly at 00:19.", "inference": "A lady needs a hotel that is not family-friendly."}, {"modality": "vision", "evidence": "At the same moment, the person who is speaking is a lady with a blonde hair and a pink ribbon.", "inference": "Confirm the lady who needs a hotel that is less family-friendly."}, {"modality": "vision", "evidence": "At the same moment, the lady on the right-hand side of the lady who needs a hotel that is less family-friendly has loose curly hair.", "inference": "The person with loose curly hair is on the right-hand side of the person who needs a hotel that is not family-friendly."}]}
{"id": "video_417__q0000", "video_id": "video_417", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_417_t5s.mp4"], "extra_info": {"question": "While the flute player was singing, what were the Santur player, the female Bushehri Deyre player, and the female Damam player doing?"}, "reward_model_ground_truth": {"multi_choice": ["A: Santur and Bushehri Deyre players were clapping and singing; Damam player was singing and playing Damam", "B: Santur and Bushehri Deyre players were clapping only; Damam player was singing and playing Damam", "C: Santur player was clapping and singing; Bushehri Deyre player was clapping; Damam player was singing and clapping", "D: Santur player was playing Santur softly; Bushehri Deyre player was clapping and singing; Damam player was singing and playing Damam"], "correct_option": "A"}, "answer_text": "Santur and Bushehri Deyre players were clapping and singing; Damam player was singing and playing Damam", "reasoning_steps": [{"modality": "audio", "evidence": "The flute player starts singing.", "inference": "Marks the start of the observation window."}, {"modality": "vision", "evidence": "The Santur player is clapping and singing.", "inference": "Identifies Santur player's actions."}, {"modality": "vision", "evidence": "The female Bushehri Deyre player is clapping and singing.", "inference": "Identifies Bushehri Deyre player's actions."}, {"modality": "vision", "evidence": "The female Damam player is playing the Damam while singing.", "inference": "Identifies Damam player's actions."}, {"modality": "audio", "evidence": "Singing from all three is heard, along with Damam beats.", "inference": "Confirms and cross-checks with the visual cues."}]}
{"id": "video_418__q0000", "video_id": "video_418", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_418_t5s.mp4"], "extra_info": {"question": "Why did Sheldon say, 'it does hurt.'?"}, "reward_model_ground_truth": {"multi_choice": ["A Because he stayed up very late last night.", "B Because his head actually hit the wall when another man carried him to bed.", "C Because his head hit the wall in his dream when another man carried him to bed.", "D Because he drank a lot of coffee last night."], "correct_option": "B"}, "answer_text": "Because his head actually hit the wall when another man carried him to bed.", "reasoning_steps": [{"modality": "vision", "evidence": "The situation showed that Sheldon was forced to stay up late to work.", "inference": "Sheldon felt unhappy and exhausted about being forced to work in the lab by his friends."}, {"modality": "audio", "evidence": "At 1:21, Sheldon jokingly said he had a dream where his friend became a tyrant and forced him to work.", "inference": "Sheldon expressed his dissatisfaction about last night jokingly."}, {"modality": "audio", "evidence": "His friend had a dream where Sheldon's head hit the wall when he carried Sheldon to bed.", "inference": "His friend might be telling the fact jokingly."}, {"modality": "audio", "evidence": "At 1:40, Sheldon said painfully, 'It does hurt!'", "inference": "Sheldon's head actually hurt, which shows this wasn't a dream, but a fact."}]}
{"id": "video_419__q0000", "video_id": "video_419", "qid": 0, "question_type": "background&music understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_419_t5s.mp4"], "extra_info": {"question": "What does the drummer's count-in suggest about the band's understanding of musical time?"}, "reward_model_ground_truth": {"multi_choice": ["A. The band is about to play a song with a purposefully missing third beat for artistic effect.", "B. The drummer is using an unconventional counting method to establish a unique and complex rhythm.", "C. The drummer has made an error in counting the typical four-beat measure before starting the song.", "D. The band intends to play a piece in an odd time signature, indicated by the irregular count."], "correct_option": "C"}, "answer_text": "The drummer has made an error in counting the typical four-beat measure before starting the song.", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:07, a drummer and a band appear in the video.", "inference": "To check the drummer's next moves."}, {"modality": "vision", "evidence": "At 2:09, he puts up the drum sticks and the other two are waiting.", "inference": "To check the audio to see if he's counting."}, {"modality": "audio", "evidence": "At the same time, the drummer starts to count:1, 2, 3, 4", "inference": "So the band mistakenly thinks that they should be counting numbers from one to four before starting the song instead of counting the typical four-beat measure."}]}
{"id": "video_419__q0001", "video_id": "video_419", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_419_t5s.mp4"], "extra_info": {"question": "Why does this man need to be burned?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because he questioned the shape of the Earth.", "B. Because the jam on the bread fell on the floor.", "C. Because he rebelled against religious rule.", "D. Because he contradicted the leader here."], "correct_option": "B"}, "answer_text": "Because the jam on the bread fell on the floor.", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:24, religious leader said 'The toast shall decide your fate. Ha ha! Guilty! Burn him!'.", "inference": "It indicates the situation of the toast is the reason why this man need to be burned."}, {"modality": "vision", "evidence": "At 1:27, after the jam on the bread fell to the ground, the man was burned.", "inference": "Because the jam on the bread fell on the floor."}]}
{"id": "video_419__q0002", "video_id": "video_419", "qid": 2, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_419_t5s.mp4"], "extra_info": {"question": "If scientific principles were to be suddenly embraced in this depicted 'World Without Science', which of the following scenarios would experience the most significant and immediate positive transformation?"}, "reward_model_ground_truth": {"multi_choice": ["A. Reliable, modern communication systems would quickly replace primitive methods.", "B. Healthcare practices would become effective, based on accurate diagnosis and surgical techniques.", "C. Societal decision-making would shift from superstitious beliefs to evidence-based reasoning.", "D. The arts and entertainment would evolve with more complex and diverse forms of expression."], "correct_option": "C"}, "answer_text": "Societal decision-making would shift from superstitious beliefs to evidence-based reasoning.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:35, we can see a doctor character.", "inference": "To check how does he makes decisions."}, {"modality": "vision", "evidence": "At 0:36, the close-up of the video shows that the doctor is working with Tarot.", "inference": "The 1st decision making way based on superstitutions."}, {"modality": "vision", "evidence": "At 1:05， people with torches gathered around a man tied up in a polar and leader is ready to make a decision.", "inference": "To check how does the leader makes decisions."}, {"modality": "audio", "evidence": "At 1:20, a man says: the toast shall decide your fate.", "inference": "The 2nd decision making way based on superstitutions."}, {"modality": "vision", "evidence": "At 1:40, a lady is sitting behind a table says BREAKING NEWS!", "inference": "To see how do they broadcast the news."}, {"modality": "audio", "evidence": "At 1:45, a lady says: And now, the weather.", "inference": "To see how do they predict the weather conditions."}, {"modality": "vision", "evidence": "At 1:50, a man says: My knees hurt, it might rain.", "inference": "The 3rd decision making way based on superstitutions, so the sudden acceptance of science will significantly change their way of making decisions."}]}
{"id": "video_42__q0000", "video_id": "video_42", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_42_t5s.mp4"], "extra_info": {"question": "When the anchor said 'Breathe has made this top lane very volatile', who got the third kill?"}, "reward_model_ground_truth": {"multi_choice": ["A.Breathe.", "B.Tian.", "C.Light.", "D.369."], "correct_option": "C"}, "answer_text": "Light.", "reasoning_steps": [{"modality": "audio", "evidence": "Tthe anchor says 'Breathe has made this top lane very volatile' at 0:21.", "inference": "From this time we can start to search the answer."}, {"modality": "vision", "evidence": "Breathe gets the first kill of 369 at 0:32.", "inference": "Breathe gets the first kill."}, {"modality": "vision", "evidence": "Tian gets the second kill of Kanavi at 1:12.", "inference": "Tian get the secodn kill."}, {"modality": "vision", "evidence": "Light gets the third kill of Hang at 1:19.", "inference": "Light gets the third kill."}]}
{"id": "video_42__q0001", "video_id": "video_42", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_42_t5s.mp4"], "extra_info": {"question": "After the anchor said 'But Breathe is just taking a huge chunk', whose kill was the third one that JackeyLove got?"}, "reward_model_ground_truth": {"multi_choice": ["A.Breathe.", "B.Tian.", "C.Light.", "D.Xiaohu."], "correct_option": "D"}, "answer_text": "Xiaohu.", "reasoning_steps": [{"modality": "audio", "evidence": "The anchor says 'But Breathe is just taking a huge chunk' at 7:32.", "inference": "This is the time point from which we can search for the answer."}, {"modality": "vision", "evidence": "After that time, JackeyLove gets the first kill of Light.", "inference": "Light's kill is the first one that JackeyLove gets."}, {"modality": "vision", "evidence": "After that time, JackeyLove gets the second kill of Tian.", "inference": "Tian's kill is the first one that JackeyLove gets."}, {"modality": "vision", "evidence": "After that time,JackeyLove gets the third kill of Xiaohu.", "inference": "Xiaohu's kill is the third one that JackeyLove gets."}]}
{"id": "video_420__q0000", "video_id": "video_420", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_420_t5s.mp4"], "extra_info": {"question": "Which instrument is audible but not visible in this video?"}, "reward_model_ground_truth": {"multi_choice": ["A. Axatse", "B. Singing", "C. Ewe drum", "D. Trumpet"], "correct_option": "D"}, "answer_text": "Trumpet", "reasoning_steps": [{"modality": "audio", "evidence": "People are signing.", "inference": "There is signing in the video."}, {"modality": "vision", "evidence": "The song contains some kind of drum.", "inference": "There are Ewe drums in the video."}, {"modality": "vision", "evidence": "The song contains Axatses.", "inference": "There are Axatses in the video."}]}
{"id": "video_421__q0000", "video_id": "video_421", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_421_t5s.mp4"], "extra_info": {"question": "If the woman wearing black and white Nike sneakers were to fully adopt her friend's stated preferences, which social media platform would she choose to 'marry'?"}, "reward_model_ground_truth": {"multi_choice": ["A. TikTok", "B. YouTube Shorts", "C. Instagram", "D. She did not state a preference."], "correct_option": "C"}, "answer_text": "Instagram", "reasoning_steps": [{"modality": "vision", "evidence": "The woman wearing black and white Nike sneakers stands on the far left of the group of three, visible starting at 00:13.", "inference": "Identify the person mentioned in the question is the left one"}, {"modality": "vision", "evidence": "The woman at the right of the group, wearing khaki pants and holding a cup with an orange drink answered the question", "inference": "Mark the point that needs attention to focus on."}, {"modality": "audio", "evidence": "She says, 'Marry Instagram, f*ck TikTok, kill the other one... YouTube Shorts'.", "inference": "Point out that she chose to marry Instagram."}, {"modality": "vision", "evidence": "The woman wearing Nike sneakers nods after her friend saying.", "inference": "Point out her agreement of 'marrying' is Instagram as well."}]}
{"id": "video_421__q0001", "video_id": "video_421", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_421_t5s.mp4"], "extra_info": {"question": "Which sequence of events accurately describes the interviewer's progression through the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. The interviewer introduces 'City Social', then asks a single person about TikTok purchases, and finally asks a group a 'Marry/Kill' question about social media platforms.", "B. The interviewer asks a group a 'Marry/Kill' question, then introduces 'City Social', and finally asks a single person about TikTok purchases.", "C. The interviewer asks a single person about TikTok purchases, then introduces 'City Social', and finally asks a group a 'Marry/Kill' question.", "D. The interviewer introduces 'City Social', then asks a group a 'Marry/Kill' question, and finally asks a single person about TikTok purchases."], "correct_option": "A"}, "answer_text": "The interviewer introduces 'City Social', then asks a single person about TikTok purchases, and finally asks a group a 'Marry/Kill' question about social media platforms.", "reasoning_steps": [{"modality": "audio", "evidence": "At the very beginning of the video, the interviewer introduces'City Social'.", "inference": "The event that the interviewer introduces'City Social' is firstly happened."}, {"modality": "audio", "evidence": "At 0:07, the interviewer asks 'Have you ever bought anything because of TikTok? '.", "inference": "The interviewer is probably asking the interviewee a question about TikTok purchases."}, {"modality": "vision", "evidence": "At 0:07, a single people is asked by the interviewer.", "inference": "The event that the interviewer asks a single person about TikTok purchases is secondly happened."}, {"modality": "audio", "evidence": "At 0:13, the interviewer asks 'Mary or kill, TikTok... '.", "inference": "At this point, the interviewer is asking another group of interviewees."}, {"modality": "vision", "evidence": "At 0:17, a group is asked by the interviewer.", "inference": "The event that the interviewer asks a group a 'Marry/Kill' question about social media platforms is thirdly happened."}]}
{"id": "video_422__q0000", "video_id": "video_422", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_422_t5s.mp4"], "extra_info": {"question": "According to the video, which single category of harvested items had the greatest number of distinct varieties mentioned or clearly shown?"}, "reward_model_ground_truth": {"multi_choice": ["A. Eggplants", "B. Peppers", "C. Berries", "D. Tomatoes"], "correct_option": "C"}, "answer_text": " Berries", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:56, the speaker said 'This is everything we got today.'.", "inference": "At this point, the camera will most likely include all the vegetables and fruits and so on that she had picked."}, {"modality": "vision", "evidence": "At 0:56, series of picked fruits, vegetables, and the like are shown in the scene.", "inference": "Identify the single category of harvested items had the greatest number of distinct varieties."}, {"modality": "vision", "evidence": "At 0:56, three different types of berries are shown in the scene.", "inference": "So the berries had the greatest number of distinct varieties mentioned."}]}
{"id": "video_423__q0000", "video_id": "video_423", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_423_t5s.mp4"], "extra_info": {"question": "If the challenge of blocking Cristiano Ronaldo's shots had continued with an increasing number of defenders, what outcome was explicitly hinted at as a reason to stop the game?"}, "reward_model_ground_truth": {"multi_choice": ["A. Ronaldo's shots would become too easy to block due to the sheer number of bodies.", "B. The defenders would risk serious injury from continuously being hit by powerful shots.", "C. The ball would inevitably get trapped within the dense wall of players, making scoring impossible.", "D. The challenge would become boring for the audience due to the lack of successful shots."], "correct_option": "B"}, "answer_text": "The defenders would risk serious injury from continuously being hit by powerful shots.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:11, a man said'That hurts!'.", "inference": "It indicates that Ronaldo injured a defender while taking a shot."}, {"modality": "vision", "evidence": "At 0:11, one defender pointed at another defender's hand.", "inference": "One defend hurted his hand."}, {"modality": "audio", "evidence": "At 0:19, this defender said 'We're calling it before someone breaks another.'.", "inference": "So the game was stopped to prevent the defenders from getting injured."}]}
{"id": "video_423__q0001", "video_id": "video_423", "qid": 1, "question_type": "summarization", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_423_t5s.mp4"], "extra_info": {"question": "How did the number of people attempting to stop Ronaldo's shots change throughout the video, and what was the outcome of each attempt?"}, "reward_model_ground_truth": {"multi_choice": ["A. The number of defenders increased from one to four, and Ronaldo scored on every single attempt.", "B. The number of defenders increased from one to four, with Ronaldo being successfully blocked only by two defenders.", "C. The number of defenders decreased from four to one, and Ronaldo scored only against the single defender.", "D. The number of defenders remained constant at three, and Ronaldo was blocked on two of his attempts."], "correct_option": "A"}, "answer_text": "The number of defenders increased from one to four, and Ronaldo scored on every single attempt.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:04, there was one defender in the scene.", "inference": "At the beginning, the number of defenders is one."}, {"modality": "vision", "evidence": "At 0:17, there were four defenders in the scene.", "inference": "In the end, the number of defenders are four. So the number of defenders increased from one to four."}, {"modality": "vision", "evidence": "At 0:05, Ronaldo successfully scored a goal.", "inference": "The first defender did not block Ronaldo."}, {"modality": "audio", "evidence": "At 0:05, A goal alert sound went off.", "inference": "Ronaldo scored on the first attempt."}, {"modality": "vision", "evidence": "At 0:06, Ronaldo successfully scored a goal again.", "inference": "The two defenders did not block Ronaldo."}, {"modality": "audio", "evidence": "At 0:06, A goal alert sound went off again.", "inference": "Ronaldo scored on the second attempt."}, {"modality": "vision", "evidence": "At 0:10, Ronaldo successfully scored a goal again.", "inference": "The three defenders did not block Ronaldo."}, {"modality": "audio", "evidence": "At 0:10, A goal alert sound went off again.", "inference": "Ronaldo scored on the third attempt."}, {"modality": "vision", "evidence": "At 0:18, Ronaldo successfully scored a goal again.", "inference": "The four defenders did not block Ronaldo."}, {"modality": "audio", "evidence": "At 0:18, A goal alert sound went off again.", "inference": "Ronaldo scored on the fourth attempt. So Ronaldo scored on every single attempt."}]}
{"id": "video_424__q0000", "video_id": "video_424", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_424_t5s.mp4"], "extra_info": {"question": "What book fell onto who felt the tool complicated?"}, "reward_model_ground_truth": {"multi_choice": ["A. David and Goliath", "B. The Bible", "C. Walden", "D. latest Gladwell"], "correct_option": "C"}, "answer_text": "Walden", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:07, the woman mentioned the tool is awfully complicated.", "inference": "The woman is the people feeling the tool complicated."}, {"modality": "vision", "evidence": "At 1:25, after she is ejected from the toilet and lands on her bed, the book 'Walden' falls onto her back.", "inference": "Walden fell onto the woman's back."}]}
{"id": "video_424__q0001", "video_id": "video_424", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_424_t5s.mp4"], "extra_info": {"question": "How many dummies finally fell onto the bed?"}, "reward_model_ground_truth": {"multi_choice": ["A. 0", "B. 1", "C. 2", "D. 3"], "correct_option": "B"}, "answer_text": "1", "reasoning_steps": [{"modality": "audio", "evidence": "At 02:13, the narrator says, 'Our team of engineers guarantees that the Toilet Death Ejector is mostly accurate...'", "inference": "It's the point to show the effect of the Toilet Death Ejector through dummies."}, {"modality": "vision", "evidence": "At 1:38, the first dommy was shown and fell to the ground.", "inference": "There is 1 dummy that did not fell onto the bed."}, {"modality": "vision", "evidence": "At 1:39, the second dommy was shown and fell onto the ground.", "inference": "There is 1 dummy that finally fell onto the bed."}, {"modality": "vision", "evidence": "At 1:41, the third dommy was shown and fell onto the ground.", "inference": "There is 1 dummy that did not fell onto the bed. So there is only one dummy fell onto the bed."}]}
{"id": "video_425__q0000", "video_id": "video_425", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_425_t5s.mp4"], "extra_info": {"question": "What is the relationship between the two persons"}, "reward_model_ground_truth": {"multi_choice": ["A. The speaker is Jingjing, and they are colleagues.", "B. The person in the blue jacket is Nana, and they are colleagues.", "C. The person in the blue jacket is Jingjing, and they are classmates.", "D. The speaker is Nana, and they are classmates."], "correct_option": "B"}, "answer_text": "They are colleagues, and the person in the blue jacket is named Nana.", "reasoning_steps": [{"modality": "vision", "evidence": "The tiger is jumping to the woman.", "inference": "The tiger is jumping."}, {"modality": "audio", "evidence": "The people filming the video called that Jingjing is jumping to Nana.", "inference": "The person in the blue jacket is Nana."}, {"modality": "audio", "evidence": "The people filming the video knows the name of the tiger and woman.", "inference": "They are colleagues."}]}
{"id": "video_426__q0000", "video_id": "video_426", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_426_t5s.mp4"], "extra_info": {"question": "What is Charlie doing when she mentions the F5 and A4 planes?"}, "reward_model_ground_truth": {"multi_choice": ["A. She is counting on her fingers.", "B. She crosses her arms over her chest.", "C. She points somewhere in the classroom.", "D. She puts her hands in her pockets."], "correct_option": "A"}, "answer_text": "She is counting on her fingers.", "reasoning_steps": [{"modality": "audio", "evidence": " At 0:40, Charlie says 'F5, A4, and Mig 23'.", "inference": "Charlie is saying the names of these planes."}, {"modality": "vision", "evidence": "While saying the names of these planes, she is counting on her fingers", "inference": "Charlie is saying the names of these planes."}]}
{"id": "video_429__q0000", "video_id": "video_429", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_429_t5s.mp4"], "extra_info": {"question": "How many times did Ru call Chad as a women?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 5"], "correct_option": "A"}, "answer_text": "2", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:42, the person with a blank expression wearing a coat says his name is 'Chad'.", "inference": "This identifies who Chad is in the video."}, {"modality": "audio", "evidence": "RuPaul says 'All she needs is a crown.' at the timestamp 1:14.", "inference": "This is the first time Ru call Chad as a women."}, {"modality": "audio", "evidence": "RuPaul says 'Work it, girl!' at the timestamp 3:07.", "inference": "This is the second time Ru call Chad as a women."}]}
{"id": "video_43__q0000", "video_id": "video_43", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_43_t5s.mp4"], "extra_info": {"question": "What would happen if Peter didn't continue to interrupt Steven's spell-casting after the third interruption?"}, "reward_model_ground_truth": {"multi_choice": ["A.The whole world will forget him.", "B.The whole world will forget him except his girlfriend.", "C.The whole world will forget him except his girlfriend and Ned.", "D.The whole world will forget him except his girlfriend、Ned and his aunt May."], "correct_option": "D"}, "answer_text": "The whole world will forget him except his girlfriend、Ned and his aunt May.", "reasoning_steps": [{"modality": "vision", "evidence": "at 5:00,there is a scene of casting the spell.", "inference": "locate this point."}, {"modality": "audio", "evidence": "Steven cast a spell to make the whole world forget Peter at 4:40.", "inference": "it happened firstly."}, {"modality": "audio", "evidence": "after Peter's first interruption, Steven made sure that Peter's girlfriend wouldn't forget him.", "inference": "his girlfriend wouldn't forget him."}, {"modality": "audio", "evidence": "after Peter's second interruption, Steven made sure that Peter's girlfriend and Ned wouldn't forget him.", "inference": "Peter's girlfriend and Ned wouldn't forget him."}, {"modality": "audio", "evidence": "after Peter's third interruption, Steven made sure that Peter's girlfriend、Ned and hist aunt May wouldn't forget him.", "inference": "Peter's girlfriend、Ned and hist aunt May wouldn't forget him."}]}
{"id": "video_43__q0001", "video_id": "video_43", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_43_t5s.mp4"], "extra_info": {"question": "How many times did the man shoveling snow shovel before Peter raised his hand after entering the door?"}, "reward_model_ground_truth": {"multi_choice": ["A.4", "B.5", "C.6", "D.7"], "correct_option": "C"}, "answer_text": "6", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:34, the man said 'Peter,...'.", "inference": "get the person."}, {"modality": "vision", "evidence": "the man who wears black jacket is Peter.", "inference": "identify Peter."}, {"modality": "vision", "evidence": "Peter raised his hand after entering the door at 0:42.", "inference": "locate this point."}, {"modality": "audio", "evidence": "There were six sounds of shoveling snow", "inference": "6."}]}
{"id": "video_43__q0002", "video_id": "video_43", "qid": 2, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_43_t5s.mp4"], "extra_info": {"question": "How many light bulbs were there in the lamp behind Steven when he said 'we've used it for a lot less'?"}, "reward_model_ground_truth": {"multi_choice": ["A.4", "B.5", "C.6", "D.7"], "correct_option": "B"}, "answer_text": "5", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:45, the man said 'Steven,...'.", "inference": "get the person."}, {"modality": "vision", "evidence": "the man who wears black and red jacket is Steven.", "inference": "identify Steven."}, {"modality": "audio", "evidence": "at 3:15, Steven said 'we've used it for a lot less'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "There were 5 light bulbs behind Steven.", "inference": "5."}]}
{"id": "video_430__q0000", "video_id": "video_430", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_430_t5s.mp4"], "extra_info": {"question": "What action happens in the scene when the first instrumental section (ignoring the introduction) starts?"}, "reward_model_ground_truth": {"multi_choice": ["A. The woman stands up.", "B. The woman touches the man.", "C. The man walks towards the woman.", "D. The woman walks away from the man."], "correct_option": "C"}, "answer_text": "The man walks towards the woman.", "reasoning_steps": [{"modality": "audio", "evidence": "Some short segments without voice are played at the beginning", "inference": "Likely not the full instrumental section in question"}, {"modality": "audio", "evidence": "After singing, there is a clear instrumental section without singing", "inference": "This is the section in question."}, {"modality": "vision", "evidence": "At 3:03, first the man walks towards the woman, and second the woman walks away from the man.", "inference": "The first action is the man walking towards the woman."}]}
{"id": "video_431__q0000", "video_id": "video_431", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_431_t5s.mp4"], "extra_info": {"question": "What disease might the protagonist in the video have?"}, "reward_model_ground_truth": {"multi_choice": ["A. Erectile dysfunction", "B. Depression", "C. Diabetes", "D. Herpes"], "correct_option": "D"}, "answer_text": "Herpes", "reasoning_steps": [{"modality": "audio", "evidence": "The video mentions at 0:59 that winning a carnival game for the first time is a possible sign of being infected with herpes.", "inference": "The ads shows signs that you could be infected by herpes."}, {"modality": "vision", "evidence": "At 1:53 seconds in the video, the words' Herpes Stopper 'appeared.", "inference": "Sets the point to observe the protagnist's reaction."}, {"modality": "vision", "evidence": "At the same moment, the protagonist's facial expressions in the video are all somewhat frightened。", "inference": "The protagonist may have contracted herpes."}]}
{"id": "video_431__q0001", "video_id": "video_431", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_431_t5s.mp4"], "extra_info": {"question": "What disease might the protagonist in the video have?"}, "reward_model_ground_truth": {"multi_choice": ["A. Erectile dysfunction", "B. Depression", "C. Diabetes", "D. Herpes"], "correct_option": "D"}, "answer_text": "Herpes", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:59 that winning a carnival game for the first time is also a possible sign of being infected with herpes.", "inference": "The ads shows more signs that a person could be infected by Herpes."}, {"modality": "vision", "evidence": "At 1:53 seconds in the video, the words' Herpes Stopper 'appeared.", "inference": "Sets the point to check the protagnist's reactions."}, {"modality": "vision", "evidence": "At the same moment, the protagonist's facial expressions in the video are all somewhat frightened.", "inference": "The protagonist have contracted herpes."}]}
{"id": "video_432__q0000", "video_id": "video_432", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_432_t5s.mp4"], "extra_info": {"question": "What is the core problematic effect of stereotypes that necessitates their avoidance in negotiation?"}, "reward_model_ground_truth": {"multi_choice": ["A. They simplify complex realities, leading to a comfortable but ultimately false foundation for negotiation strategy.", "B. They make it impossible to identify regional cultural differences, thereby impeding effective cross-cultural communication.", "C. They lead to an overemphasis on national cultural traits, causing negotiators to ignore professional cultural identities.", "D. They fail to account for the unique individual characteristics of negotiators, resulting in rigid and ineffective approaches."], "correct_option": "A"}, "answer_text": "They simplify complex realities, leading to a comfortable but ultimately false foundation for negotiation strategy.", "reasoning_steps": [{"modality": "vision", "evidence": "At 3:36, the text 'Avoid stereotypes' was shown on the right side of the video screen.", "inference": "The video started to focus on things regarding stereotypes at that moment."}, {"modality": "audio", "evidence": "At 3:43, the man in the video talked about stereotypes 'simplify a reality that is much more complex'.", "inference": "According to his speech, it can be inferred that stereotypes simplify complex realities."}, {"modality": "audio", "evidence": "At 3:45, the man said, 'This simplistic reductionalism provides a comfortable starting point for elaborating the strategy.'", "inference": "According to his speech, it can be inferred that stereotypes provide a comfortable foundation for the strategy."}, {"modality": "audio", "evidence": "At 3:59, the man said the starting point is constructed on a false foundation.", "inference": "According to his speech, it can be inferred that stereotypes may lead to false foundation for the strategy. And Based on the above inferences, we know that stereotypes simplify complex realities, leading to a comfortable but ultimately false foundation for negotiation strategy."}]}
{"id": "video_432__q0001", "video_id": "video_432", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_432_t5s.mp4"], "extra_info": {"question": "What was he talking about when he rubbed his hands for the first time?"}, "reward_model_ground_truth": {"multi_choice": ["A. Why it is important to treat ecological fallacy cautiously.", "B. Why it is important to also consider professional culture.", "C. An example showing lawyers' familiarity with conflicts.", "D. Every negotiator belonging to multiple cultural circles."], "correct_option": "A"}, "answer_text": "Why it is important to treat ecological fallacy cautiously.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:35, the man raised his hands and rubbed them for the first time.", "inference": "It confirms the timeline that the man rubbed his hands mentioned in the question."}, {"modality": "audio", "evidence": "At 0:35, the man says 'Why it is important to treat ecological fallacy cautiously.'", "inference": "This indicates the answer to the question."}]}
{"id": "video_433__q0000", "video_id": "video_433", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_433_t5s.mp4"], "extra_info": {"question": "What is the value of the part with horsetail hair?"}, "reward_model_ground_truth": {"multi_choice": ["A.$10,000", "B.$35,000", "C.$60,000", "D.$600,000"], "correct_option": "D"}, "answer_text": "$600,000", "reasoning_steps": [{"modality": "aduio", "evidence": "The middle section contains horsetail hair."}, {"modality": "vision", "evidence": "The middle section is valued at $600,000.", "inference": "The answer is $600,000."}]}
{"id": "video_434__q0000", "video_id": "video_434", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_434_t5s.mp4"], "extra_info": {"question": "How many drummers appear in the scene at the same time?"}, "reward_model_ground_truth": {"multi_choice": ["A.1", "B.2", "C.3", "D.4"], "correct_option": "D"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:31, drum sounds were heard.", "inference": "Someone is drumming."}, {"modality": "vision,", "evidence": "At 0:30, count the number of drummers who appeared in the frame.", "inference": "There are four drummers appearing in the scene at the same time."}]}
{"id": "video_435__q0000", "video_id": "video_435", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_435_t5s.mp4"], "extra_info": {"question": "What is the name of the creature who is the same status as the spider?"}, "reward_model_ground_truth": {"multi_choice": ["A. Roxy", "B. Michus", "C. Johnny", "D. Bolty"], "correct_option": "B"}, "answer_text": "Michus", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:26, the girl says 'There you are Roxy. Come here and I'll feed you.'", "inference": "The spider is the girl's pet and Roxy is its name."}, {"modality": "vision", "evidence": "At 1:29, Cat sits next to the spider.", "inference": "Cat is also the pet."}, {"modality": "vision", "evidence": "At 1:29, there is another bowl written with 'Michus'.", "inference": "Michus is the name of the cat."}]}
{"id": "video_436__q0000", "video_id": "video_436", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_436_t5s.mp4"], "extra_info": {"question": "How many different sounds appear when using the red sauce?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 5"], "correct_option": "B"}, "answer_text": "3", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:11, red sauce appears in the video.", "inference": "This indicates the position of red sauce."}, {"modality": "audio", "evidence": "At 0:11 in the video, the traditional tomato ketchup bottle, fart-like.", "inference": "This is the first sound."}, {"modality": "audio", "evidence": "At 0:58 in the video,Heinz Relax ketchup, a relaxing sigh.", "inference": "This is the second sound."}, {"modality": "audio", "evidence": "At 1:23 in the video, glass bottle of tomato ketchup, 'uh uh uh' sounds.", "inference": "This is the third sound."}]}
{"id": "video_438__q0000", "video_id": "video_438", "qid": 0, "question_type": "attribute comparison", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_438_t5s.mp4"], "extra_info": {"question": "The flag on the boy's bicycle for is the same color as whose hair?"}, "reward_model_ground_truth": {"multi_choice": ["A Linda", "B Stotch", "C The person in the passenger seat when Stotch is driving", "D The person in the back seat when Stotch is driving"], "correct_option": "A"}, "answer_text": "Linda", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:01, the boy's bicycle is shown.", "inference": "The flag on the boy's bicycle is yellow."}, {"modality": "vision", "evidence": "At 0:28, the boy's mother's hair color is yellow.", "inference": "The mother's hair color is the same as the flag on the boy's bicycle."}, {"modality": "audio", "evidence": "At 0:53, the father calls mother Linda .", "inference": "Linda's hair is the same color of the flag on the boy's bicycle."}]}
{"id": "video_439__q0000", "video_id": "video_439", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_439_t5s.mp4"], "extra_info": {"question": "Where is the security guard?"}, "reward_model_ground_truth": {"multi_choice": ["A. They are shot by the robber.", "B. They are controled by the robber.", "C. They are on the ground.", "D. They are busy with being real."], "correct_option": "D"}, "answer_text": "They are busy with being real.", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:53, a woman said' What the hell is security?'.", "inference": "The security guard were not found when the robbers came."}, {"modality": "vision", "evidence": "The security guard is shown taking a selfie for the BeReal app with another person, not awaring the robbers.", "inference": "The security guard are busy with being real."}]}
{"id": "video_44__q0000", "video_id": "video_44", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_44_t5s.mp4"], "extra_info": {"question": "Before the anchor said, 'IG is so far behind in kills, how come they're playing like they have the advantage?', who was the second to last person to die?"}, "reward_model_ground_truth": {"multi_choice": ["A.Bin.", "B.Knight.", "C.On.", "D.Wei."], "correct_option": "A"}, "answer_text": "Bin.", "reasoning_steps": [{"modality": "audio", "evidence": "At 04:37, the anchor said, 'IG is so far behind in kills, how come they're playing like they have the advantage?'", "inference": "the anchor mentioned the sentence at 04:37."}, {"modality": "vision", "evidence": "At 04:21, Wei died.", "inference": "Before the anchor mentioned the sentence, the first to last person to die was Wei."}, {"modality": "vision", "evidence": "At 04:09, Bin died.", "inference": "Before the anchor mentioned the sentence, the second to last person to die was Bin."}]}
{"id": "video_440__q0000", "video_id": "video_440", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_440_t5s.mp4"], "extra_info": {"question": "From the moment the Daf starts playing until the male background singers join, what sequence of hand movements does the lead singer perform?"}, "reward_model_ground_truth": {"multi_choice": ["A: Keeps palms together, claps 7 times, keeps palms together, raises one hand, then raises both hands", "B: Raises one hand, claps 7 times, raises one hand again, keeps palms together, then raises both hands", "C: Keeps palms together, claps 8 times, keeps palms together, raises one hand, then raises both hands", "D: Raises one hand, claps 8 times, raises one hand again, keeps palms together, then raises both hands"], "correct_option": "D"}, "answer_text": "Raises one hand, claps 8 times, raises one hand again, keeps palms together, then raises both hands", "reasoning_steps": [{"modality": "audio", "evidence": "First Daf sound is heard.", "inference": "Marks the start of the observation window."}, {"modality": "vision", "evidence": "Singer raises one hand.", "inference": "First gesture."}, {"modality": "vision", "evidence": "Singer claps exactly 8 times.", "inference": "Confirms the clap count."}, {"modality": "vision", "evidence": "Singer raises one hand again.", "inference": "Third gesture."}, {"modality": "vision", "evidence": "Singer keeps palms together.", "inference": "Fourth gesture."}, {"modality": "vision", "evidence": "Singer raises both hands.", "inference": "Final gesture before background. singers start"}, {"modality": "audio", "evidence": "Male background singers start.", "inference": "Marks the end of the observation window."}]}
{"id": "video_441__q0000", "video_id": "video_441", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_441_t5s.mp4"], "extra_info": {"question": "Before the music changes, how many times does the Bedrock Player character visibly die throughout the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1 times", "B. 2 times", "C. 3 times", "D. 4 times"], "correct_option": "A"}, "answer_text": "1 times", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:12, the music changed.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "At 0:09, Bedrock falls and the subtitle shows: YOU DIED!", "inference": "The 1st time Bedrock died."}]}
{"id": "video_441__q0001", "video_id": "video_441", "qid": 1, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_441_t5s.mp4"], "extra_info": {"question": "When a Bedrock player starts taking damage in water, how many damage - taking sounds are emitted in total?"}, "reward_model_ground_truth": {"multi_choice": ["A. 3.", "B. 4.", "C. 5.", "D. 6."], "correct_option": "D"}, "answer_text": "6.", "reasoning_steps": [{"modality": "vision", "evidence": "the bedrock player starts taking damage in water at 00:13.", "inference": "the sound of the damage in water start after 00:13."}, {"modality": "audio", "evidence": "The sound of the first damage-taking appeared at 0:13", "inference": "There's one sound of the damage-taking appeared at 0:13."}, {"modality": "audio", "evidence": "The sound of the second damage-taking appeared at 0:14", "inference": "There're two sound of the damage-taking appeared at 0:14."}, {"modality": "audio", "evidence": "The sound of the third damage-taking appeared at 0:14", "inference": "There're three sound of the damage-taking appeared at 0:14."}, {"nmodality": "audio", "evidence": "The sound of the first damage-taking appeared at 0:15", "inference": "There's one sound of the damage-taking appeared at 0:15."}, {"modality": "audio", "evidence": "The sound of the second damage-taking appeared at 0:15", "inference": "There're two sound of the damage-taking appeared at 0:15, so there're totally six sound of the damage-taking appeared."}]}
{"id": "video_443__q0000", "video_id": "video_443", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_443_t5s.mp4"], "extra_info": {"question": "What is the relationship between Kate and Dale?"}, "reward_model_ground_truth": {"multi_choice": ["A. his daughter", "B. Neighbor's daughter", "C. His son's playmate", "D. His wife's illegitimate daughter"], "correct_option": "B"}, "answer_text": "Neighbor's daughter", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:12, a man said 'Kate'.", "inference": "This is the girl mentioned in the question."}, {"modality": "vision", "evidence": "At 0:12, a girl with dark brown hair is in the scene.", "inference": "Identify the person Kate."}, {"modality": "audio", "evidence": "At 1:40, another man said 'Dale'.", "inference": "This is the person mentioned in the question."}, {"modality": "vision", "evidence": "At 1:40, a man with a orange cap is shown .", "inference": "This is the man mentioned in the question."}]}
{"id": "video_444__q0000", "video_id": "video_444", "qid": 0, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_444_t5s.mp4"], "extra_info": {"question": "The person saying 'Don't pick up her calls' is the how many-th speaking female that has appeared in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "D"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:42, the girl in white clothes says 'Don't pick up her calls'.", "inference": "This identifies the position of this statement."}, {"modality": "vision", "evidence": "Before 0:42, there are 3 girls appearing in the video.", "inference": "Three female characters are visible before this timestamp."}, {"modality": "audio", "evidence": "The narrator is also a female voice.", "inference": "There are a total of 4 females including the narrator."}]}
{"id": "video_444__q0001", "video_id": "video_444", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_444_t5s.mp4"], "extra_info": {"question": "How many people died when introducing different usages of cheques?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "A"}, "answer_text": "1", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:58, the voiceover says 'buying poison'.", "inference": "It introduces the fourth usage of cheques."}, {"modality": "audio", "evidence": "The black woman says 'Just one sip and I become the head of the board'.", "inference": "the woman is poisoning the head of the board."}, {"modality": "vision", "evidence": "At 1:01, the man falls on the ground.", "inference": "He is killed by the woman."}]}
{"id": "video_445__q0000", "video_id": "video_445", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_445_t5s.mp4"], "extra_info": {"question": "When talking about pajamas, how many outlets are there in total on the wall behind the woman?"}, "reward_model_ground_truth": {"multi_choice": ["A 3", "B 4", "C 5", "D 6"], "correct_option": "D"}, "answer_text": "6", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:08, the woman says'Pajamas aren't supposed to be baggy, right?'.", "inference": "It is the point that people are talking about pajamas mentioned in the question."}, {"modality": "vision", "evidence": "At 0:21, there are 3 outlets on the right of the woman.", "inference": "On the right, three outlets are on the wall behind the woman."}, {"modality": "vision", "evidence": "At 3:01, there are 3 outlets on the left side of the woman. Therefore, adding the 3 outlets on the right, there are  6 outlets in total.", "inference": "there are 6 outlets in total on the wall behind the woman."}]}
{"id": "video_446__q0000", "video_id": "video_446", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_446_t5s.mp4"], "extra_info": {"question": "What gesture does the speaker make when she talks about the pain of inaction?"}, "reward_model_ground_truth": {"multi_choice": ["A. She makes a gesture of waving one hand upward.", "B. She holds her hands out with her palms facing up.", "C. She rests her chin on her hand, looking thoughtful.", "D. She crosses her arms over her chest."], "correct_option": "B"}, "answer_text": "She holds her hands out with her palms facing up.", "reasoning_steps": [{"modality": "audio", "evidence": "The speaker says 'the pain of inaction' with a heavy and serious tone."}, {"modality": "vision", "evidence": "Simultaneously, she holds her hands out in front of her with her palms facing up."}, {"modality": "audio", "evidence": "The speaker continues her point with a serious tone."}, {"modality": "vision", "evidence": "The gesture emphasizes the weight of her words, contrasting with other more casual gestures."}]}
{"id": "video_448__q0000", "video_id": "video_448", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_448_t5s.mp4"], "extra_info": {"question": "What does 'I have a cheeseburger in my pocket' mean?"}, "reward_model_ground_truth": {"multi_choice": ["A. Explain why all the dogs are suddenly attracted to her.", "B. Offer her date a snack because she thinks he might be hungry.", "C. Imply that her 'Pinx' underwear is as good at containing food odors as it is with her period.", "D. Make a random joke to see if her date has a similar sense of humor."], "correct_option": "A"}, "answer_text": "Explain why all the dogs are suddenly attracted to her.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:55, the woman found all the dogs were suddenly attracted to her.", "inference": "The dogs were attracted by the woman."}, {"modality": "audio", "evidence": "At 0:58, the man told the woman that the dog really liked her.", "inference": "The man is curious about why the dog likes her."}, {"modality": "audio", "evidence": "At 1:01, the woman said that she has a cheeseburger in her pocket.", "inference": "Explain why all the dogs are suddenly attracted to her."}]}
{"id": "video_449__q0000", "video_id": "video_449", "qid": 0, "question_type": "attribute comparison", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_449_t5s.mp4"], "extra_info": {"question": "What additional sounds did the best imitators add to the horse's voice compared to others?"}, "reward_model_ground_truth": {"multi_choice": ["A.Horse neighing", "B Horse chewing", "C.Horse hoofbeats", "D.Hoorse snorting"], "correct_option": "C"}, "answer_text": "Horse hoofbeats", "reasoning_steps": [{"modality": "vision", "evidence": "The upper part of the video is a picture of a horse at 00:53.", "inference": "The photographers are imitating the sound of horses."}, {"modality": "vision", "evidence": "The photographer on the far left ranks first with a score of 100 at 00:53.", "inference": "He is the best voice imitator."}, {"modality": "audio", "evidence": "The imitator on the far left made the sound of a horse's hoofbeats as it moved at 01:02.", "inference": "The sound of horse hoofbeats makes him win."}]}
{"id": "video_45__q0000", "video_id": "video_45", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_45_t5s.mp4"], "extra_info": {"question": "Why did the two judges vote for plate one at first?"}, "reward_model_ground_truth": {"multi_choice": ["A.They think Plate One has soul.", "B.They think Plate One is delicious.", "C.They think Plate Two lacks soul.", "D.They thought that Plate One was made by a Black person."], "correct_option": "D"}, "answer_text": "They thought that Plate One was made by a Black person.", "reasoning_steps": [{"modality": "vision", "evidence": "Both judges are people of color.", "inference": "maybe the video is satirizing issues of skin color."}, {"modality": "audio", "evidence": "At 1:15, the two judges said 'she'.", "inference": "maybe they pointed the woman chef."}, {"modality": "audio", "evidence": "At 1:15, the two judges said the plate one connects with the woman chef.", "inference": "They thought that Plate One was made by a woman."}, {"modality": "vision", "evidence": "the woman chef is a black person.", "inference": "They thought that Plate One was made by a Black person."}]}
{"id": "video_45__q0001", "video_id": "video_45", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_45_t5s.mp4"], "extra_info": {"question": "Why was Plate One selected by the judges?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because Plate One tastes very good.", "B.Because Plate One has a soul.", "C.Because the judges think Plate One was made by a Black person.", "D.Because the judges think Plate One was made by a White person."], "correct_option": "C"}, "answer_text": "Because the judges think Plate One was made by a Black person.", "reasoning_steps": [{"modality": "vision", "evidence": "Both judges are people of color.", "inference": "maybe the video is satirizing issues of skin color."}, {"modality": "audio", "evidence": "At 1:15, the two judges said 'she'.", "inference": "maybe they pointed the woman chef."}, {"modality": "audio", "evidence": "At 1:15, the two judges said the plate one connects with the woman chef.", "inference": "They thought that Plate One was made by a woman."}, {"modality": "vision", "evidence": "the woman chef is a black person.", "inference": "They thought that Plate One was made by a Black person."}]}
{"id": "video_450__q0000", "video_id": "video_450", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_450_t5s.mp4"], "extra_info": {"question": "Which one is the person on the left of the one holding a drink among those who said they have no shoes?"}, "reward_model_ground_truth": {"multi_choice": ["A The first", "B The second", "C The third", "D The fourth"], "correct_option": "B"}, "answer_text": "The second", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:03, the purple bat says 'I don't have my shoes'.", "inference": "The first mentions no shoes."}, {"modality": "audio", "evidence": "Patrick says 'I don't have my shoes either'.", "inference": "Patrick is the second person who says they have no shoes."}, {"modality": "vision", "evidence": "At 0:53, the color of the character holding a drink is green.", "inference": "identify the one holding a drink"}, {"modality": "vision", "evidence": "At 0:53, the first character on his left is Patrick.", "inference": "The second person who says they have no shoes is on the left of the one holding a drink."}]}
{"id": "video_451__q0000", "video_id": "video_451", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_451_t5s.mp4"], "extra_info": {"question": "How many people tried it without shouting 'MY GOD'?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "B"}, "answer_text": "2", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:11, the main character, football player, who uses the Olay Eye Black throughout the ad.", "inference": "The football player is the first one who tried it."}, {"modality": "audio", "evidence": "He doesn't shouting 'MY GOD' throughout the ad.", "inference": "The first one who tried it without shouting 'MY GOD'."}, {"modality": "vision", "evidence": "At 0:46 in the video, another man tried it.", "inference": "He is the second one who tried it."}, {"modality": "audio", "evidence": "The second man tried the eye black and questioned its masculinity before deciding it's 'just gay enough.'", "inference": "The second on who tried it without shouting 'MY GOD'."}, {"modality": "vision", "evidence": "A man tries the 'Pore Strip Nose Tape' at 1:05 in the video.", "inference": "He is the third one who tried it."}, {"modality": "audio", "evidence": "At 1:11 in the video,the third man shout 'OH MY GOD!' after using it.", "inference": "The third man is the only one who tried it with shouting 'MY GOD', so there are 2 people tried it without shouting 'MY GOD'."}]}
{"id": "video_452__q0000", "video_id": "video_452", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_452_t5s.mp4"], "extra_info": {"question": "What is the use of bad eggs that have been detected?"}, "reward_model_ground_truth": {"multi_choice": ["A. Become fertilizer for crops.", "B. Become food for ducks.", "C. Become food for fish.", "D. Become props for video shooting."], "correct_option": "C"}, "answer_text": "Become food for fish.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:00, the arrow points to eggs.", "inference": "The arrow is indicating eggs as the subject of focus."}, {"modality": "audio", "evidence": "At 0:00, the narrator says 'Why do the Chinese throw their eggs in the water and waste them?'", "inference": "This indicates these are bad eggs being discarded."}, {"modality": "audio", "evidence": "At 0:18 seconds in the video, the narrator said that 'the bad eggs that are throwed in the water are used as fish-feed.", "inference": "Bad eggs are used as fish feed."}]}
{"id": "video_453__q0000", "video_id": "video_453", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_453_t5s.mp4"], "extra_info": {"question": "What is the price tag of the item behind the color of Stan's mother's wrist when the salesperson says 'Can I help you with anything?'"}, "reward_model_ground_truth": {"multi_choice": ["A. 59.99", "B. 22.99", "C. 54.99", "D. 47.99"], "correct_option": "B"}, "answer_text": "22.99", "reasoning_steps": [{"modality": "vision", "evidence": "When they are at home, the color of Stan's mother's wrist is red.", "inference": "We should pay attention to the price tag of the red item appearing in the supermarket."}, {"modality": "audio", "evidence": "At 1:23, the salesperson says 'Can I help you with anything?'", "inference": "Pay attention to the price tag of the red item behind them."}, {"modality": "vision", "evidence": "At 1:23, we can see the price tag of the red item is 22.99.", "inference": "The price tag of the item is 22.99."}]}
{"id": "video_454__q0000", "video_id": "video_454", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_454_t5s.mp4"], "extra_info": {"question": "When the first section that is upbeat rhythmically starts, the camera focuses on a woman. What color are most of the flowers in her head?"}, "reward_model_ground_truth": {"multi_choice": ["A. Red", "B. Orange", "C. Green", "D. Pink"], "correct_option": "B"}, "answer_text": "Orange", "reasoning_steps": [{"modality": "audio", "evidence": "The first parts of the song are slow and contemplative.", "inference": "These are not the upbeat section."}, {"modality": "audio", "evidence": "At timestamp 0:41, a more rhythmical, upbeat section starts.", "inference": "This is the section in question."}, {"modality": "vision", "evidence": "During that section, a single woman is in focus.", "inference": "This is the woman in question."}, {"modality": "vision", "evidence": "The woman is wearing several flowers on her head, with the majority being orange.", "inference": "The answer is orange."}]}
{"id": "video_455__q0000", "video_id": "video_455", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_455_t5s.mp4"], "extra_info": {"question": "In which direction is his girlfriend most likely to be relative to him?"}, "reward_model_ground_truth": {"multi_choice": ["A. Front", "B. Back", "C. Left", "D. Right"], "correct_option": "C"}, "answer_text": "Left", "reasoning_steps": [{"modality": "audio", "evidence": "At 00:41, the livestreamer's girlfriend starts to argue about whether he made a WeChat call.", "inference": "Mark the start point of an argue between them."}, {"modality": "vision", "evidence": "The live streamer turns his head to the left to respond to his girlfriend's questioning.", "inference": "Confimed that his girlfriend is on his left."}]}
{"id": "video_456__q0000", "video_id": "video_456", "qid": 0, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_456_t5s.mp4"], "extra_info": {"question": "Who is the same person that the one saying 'You know what that means' is surprised about the weight number?"}, "reward_model_ground_truth": {"multi_choice": ["A. the one with polar bear arms", "B. the one made his own heat with his big body", "C. the one with 'WHOLE DAME KEG' on his belly", "D. the one in the first class when he go travel"], "correct_option": "D"}, "answer_text": " the one in the first class when he go travel", "reasoning_steps": [{"modality": "audio", "evidence": "The person saying 'You know what that means' is the one with big wavy hair.", "inference": "This identifies the speaker with big wavy hair."}, {"modality": "vision", "evidence": "At 3:01, the person with big wavy hair is surprised about the weight of the person with curly hair.", "inference": "The person with curly hair is the one whose weight causes surprise."}, {"modality": "vision", "evidence": "This black man with curl hair in the video at 2:16 is in the airplane's first-class.", "inference": "The man is the one in the first class when he goes traveling."}]}
{"id": "video_456__q0001", "video_id": "video_456", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_456_t5s.mp4"], "extra_info": {"question": "Before one asked, how many no 'big boys' showed up?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 5"], "correct_option": "B"}, "answer_text": "3", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:18, the woman says they are looking for some 'big boys'.", "inference": "Answer will appear before 0:18."}, {"modality": "vision", "evidence": "At 0:10, one man is in the center wearing a plaid jacket and holding a cup.", "inference": "He is not 'big boys'."}, {"modality": "vision", "evidence": "At 0:10, one man wearing plaid vest is on the left.", "inference": "This man is also not 'big boys'."}, {"modality": "vision", "evidence": "At 0:10, one man wearing plaid vest is on the right.", "inference": "This man is also not 'big boys'."}]}
{"id": "video_457__q0000", "video_id": "video_457", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_457_t5s.mp4"], "extra_info": {"question": "When the video's voiceover says '10 thousand copies' ,how many characters are on the computer screen's game interface?"}, "reward_model_ground_truth": {"multi_choice": ["A 1", "B 2", "C 3", "D 4"], "correct_option": "B"}, "answer_text": "2", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:56, the voiceover mentions '10 thousand copies'.", "inference": "After this moment, the scene may show the computer screen's game interface."}, {"modality": "vision", "evidence": "At 0:59, there are 2 characters visible in the game interface at that moment.", "inference": "The answer is 2."}]}
{"id": "video_458__q0000", "video_id": "video_458", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_458_t5s.mp4"], "extra_info": {"question": "How many push-ups did the record-breaker asked 'what did you do today' complete?"}, "reward_model_ground_truth": {"multi_choice": ["A 19", "B 10", "C 12", "D 26"], "correct_option": "A"}, "answer_text": "19", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:17, the person who answered when the record-breaker said 'what did you do today.'.", "inference": "This is the record-breaker mentioned in the question."}, {"modality": "vision", "evidence": "At 0:17, the person is Jim.", "inference": "Jim is the record-breaker."}, {"modality": "audio", "evidence": "At 0:56, Jim said that he did 19 push-ups.", "inference": "So the record-breaker complete 19 push-ups."}]}
{"id": "video_46__q0000", "video_id": "video_46", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_46_t5s.mp4"], "extra_info": {"question": "When 'meow' is said in the video, how many fingers do the people in the frame have?"}, "reward_model_ground_truth": {"multi_choice": ["A.3.", "B.4.", "C.5.", "D.6."], "correct_option": "B"}, "answer_text": "4.", "reasoning_steps": [{"modality": "audio", "evidence": "mario said 'meow' at 4:36.", "inference": "get the 'meow'."}, {"modality": "vision", "evidence": "From vision, he has 4 fingers.", "inference": "4."}]}
{"id": "video_460__q0000", "video_id": "video_460", "qid": 0, "question_type": "background&music understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_460_t5s.mp4"], "extra_info": {"question": "Why does the background music suddenly appear?"}, "reward_model_ground_truth": {"multi_choice": ["A. To contrast with the previous funny footsteps", "B. To highlight that the protagonist is too fat", "C. To highlight that the protagonist is too slow", "D. To highlight that the protagonist is very tired"], "correct_option": "A"}, "answer_text": "To contrast with the previous funny footsteps", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:12, the background music starts to appear.", "inference": "The background music begins at this specific timestamp."}, {"modality": "vision", "evidence": "At 0:01 seconds in the video, the man in the video going up the mountain is very funny.", "inference": "The raw footage is very funny.So answer the question."}]}
{"id": "video_460__q0001", "video_id": "video_460", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_460_t5s.mp4"], "extra_info": {"question": "How many times does the man vocally pronounce 'Hit' while directly facing the camera at the beginning of the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2 times", "B. 3 times", "C. 4 times", "D. 5 times"], "correct_option": "A"}, "answer_text": "2 times", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:01, the man vocally pronounce 'Hit.", "inference": "To check if he's pronouncing it facing the camera."}, {"modality": "vision", "evidence": "At the same moment, he is facing the camera.", "inference": "Counted as 1 time he's saying 'Hit' while facing the camera."}, {"modality": "audio", "evidence": "At 0:02, the man vocally pronounce 'Hit again.", "inference": "To check if he's pronouncing it facing the camera."}, {"modality": "vision", "evidence": "At the same moment, he is facing the camera.", "inference": "Counted as 2 times he's saying 'Hit' while facing the camera in total."}]}
{"id": "video_460__q0002", "video_id": "video_460", "qid": 2, "question_type": "causal reasoning", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_460_t5s.mp4"], "extra_info": {"question": "Why did the music change like that when the scene changed?"}, "reward_model_ground_truth": {"multi_choice": ["A. To create a sad atmosphere.", "B. To create a heavy atmosphere.", "C. To create a free-spirited atmosphere.", "D. To create an atmosphere of anger."], "correct_option": "C"}, "answer_text": "To create a free-spirited atmosphere.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:06, the scene changed.", "inference": "At 0:06, the scene changed."}, {"modality": "audio", "evidence": "At the same moment, the music change to a more free rhythm.", "inference": "At the same moment, the music change to a more free rhythm."}]}
{"id": "video_465__q0000", "video_id": "video_465", "qid": 0, "question_type": "summarization", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_465_t5s.mp4"], "extra_info": {"question": "What place did Jimmy achieve?"}, "reward_model_ground_truth": {"multi_choice": ["A.1st", "B.2nd", "C.3rd", "D.4th"], "correct_option": "D"}, "answer_text": "4th", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:05, someone said 'Bye Jimmy!'", "inference": "The person who is about to fall is Jimmy."}, {"modality": "vision", "evidence": "At 1:07, Jimmy fell over.", "inference": "Jimmy finished."}, {"modality": "vision", "evidence": "At 1:07, there are 3 other players who were still running.", "inference": "Jimmy got the 4th."}]}
{"id": "video_466__q0000", "video_id": "video_466", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_466_t5s.mp4"], "extra_info": {"question": "Who is Mrs Geller?"}, "reward_model_ground_truth": {"multi_choice": ["A.Rachel.", "B.Monica.", "C.Ross", "D.Emily."], "correct_option": "D"}, "answer_text": "Emily", "reasoning_steps": [{"modality": "audio", "evidence": "Rachel says 'I had to stay in their hollymoon suite, when people coming up to me all the time going oh, Mrs Geller, why you cry'"}, {"modality": "vision", "evidence": "Rachel is upset but pretends to be fine in front of Ross"}, {"modality": "audio", "evidence": "Ross is try to make up for his wife, Emily."}]}
{"id": "video_468__q0000", "video_id": "video_468", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_468_t5s.mp4"], "extra_info": {"question": "What is the job of the cousin of the person who does not have the 'guidance system' patent?"}, "reward_model_ground_truth": {"multi_choice": ["A Lawyer", "B Work in a call center", "C Scientist", "D A know-it-all"], "correct_option": "B"}, "answer_text": "Work in a call center", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:20, two people show, the left is Howard in blue shirt and the right is Raj in green jacket.", "inference": "Make sure people can identify Raj as the scene changes."}, {"modality": "audio", "evidence": "At 0:25, Raj said 'Yeah, this military guy showed up at Howard's door.'.", "inference": "It indicates that Howard has the 'guidance system' patent not Raj."}, {"modality": "audio", "evidence": "At 2:08, Raj says 'My cousin does work in a call center'.", "inference": "Show the job of the cousin of Raj."}]}
{"id": "video_469__q0000", "video_id": "video_469", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_469_t5s.mp4"], "extra_info": {"question": "Which one's information on letter was not correct?"}, "reward_model_ground_truth": {"multi_choice": ["A. Daughter", "B. Mom", "C. Dad", "D. No one"], "correct_option": "D"}, "answer_text": "No one", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:28, the dad hold the letter that told the family's private information.", "inference": "Here is the point that the information is shown."}, {"modality": "audio", "evidence": "At 0:52, dad, daughter and son admitted that the information was correct.", "inference": "The information told by the letter may be all correct."}, {"modality": "audio", "evidence": "At 1:14, the mom said 'OK'.", "inference": "She is also agree what the letter said. So nobody's information on letter was not correct."}]}
{"id": "video_47__q0000", "video_id": "video_47", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_47_t5s.mp4"], "extra_info": {"question": "Among the following events, which one happened second in chronological order?"}, "reward_model_ground_truth": {"multi_choice": ["A.Wolf Parking.", "B.Wolf's resume has been destroyed.", "C.Wolf recognized Diane.", "D.Wolf thought he had gotten the job."], "correct_option": "A"}, "answer_text": "Wolf Parking.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:13,someone said 'it's Diane'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "Wolf recognized Diane.", "inference": "it happened firstly."}, {"modality": "vision", "evidence": "From vision, wolf is parking at 3:27.", "inference": "it happened secondly."}]}
{"id": "video_47__q0001", "video_id": "video_47", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_47_t5s.mp4"], "extra_info": {"question": "Among the following events, which one happened first in chronological order?"}, "reward_model_ground_truth": {"multi_choice": ["A.Wolf Parking.", "B.Wolf's resume has been destroyed.", "C.Wolf recognized Diane.", "D.Wolf thought he had gotten the job."], "correct_option": "C"}, "answer_text": "Wolf recognized Diane.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:13,someone said 'it's Diane'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "Wolf recognized Diane.", "inference": "it happened firstly."}]}
{"id": "video_470__q0000", "video_id": "video_470", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_470_t5s.mp4"], "extra_info": {"question": "Which profession does Bear express clear refusal for?"}, "reward_model_ground_truth": {"multi_choice": ["A. A chef", "B. A doctor", "C. A firefighter", "D. A sword swallower"], "correct_option": "A"}, "answer_text": "A chef", "reasoning_steps": [{"modality": "auido.", "evidence": "At 0:13, the woman on the phone said, 'We need an experienced chef.'", "inference": "The job depicted in the video is a chef."}, {"modality": "auido.", "evidence": "At 0:20, the girl says 'yeah,he is here..he is the best chef..'", "inference": "It indicates Bear ."}, {"modality": "vision", "evidence": "At 0:23, Bear explicitly waves his hand in refusal.", "inference": "The bear rejected to be a chef."}]}
{"id": "video_470__q0001", "video_id": "video_470", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_470_t5s.mp4"], "extra_info": {"question": "How do the Bear's emotions change?"}, "reward_model_ground_truth": {"multi_choice": ["A. He remains indifferent throughout the clip, showing little change.", "B. He changes from sad to enthusiastic.", "C. He initially seems bored but becomes anxious.", "D. He appears calm at first, then becomes startled and frustrated."], "correct_option": "D"}, "answer_text": "He appears calm at first, then becomes startled and frustrated.", "reasoning_steps": [{"modality": "vision", "evidence": "at the beginning , the bear was relaxing on an armchair", "inference": "the bear is calm at first"}, {"modality": "audio", "evidence": "the girl answered the call", "inference": "this triggers the emotional changing of the bear"}, {"modality": "vision", "evidence": "the bear show the expression of shock in 0:29", "inference": "the bear becomes  startled"}, {"modality": "vision", "evidence": "the bear sighed in 0:33", "inference": "the bear becomes  frustrated."}]}
{"id": "video_470__q0002", "video_id": "video_470", "qid": 2, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_470_t5s.mp4"], "extra_info": {"question": "What is Masha primarily attempting to achieve for Bear before the bear showed a surprised expression?"}, "reward_model_ground_truth": {"multi_choice": ["A. To find Bear a new friend to cheer him up from his sadness.", "B. To secure an opportunity for Bear as an experienced chef.", "C. To amuse herself by fabricating stories about Bear's hidden abilities.", "D. To book Bear a spot for a grand performance at the Moscow Circus."], "correct_option": "B"}, "answer_text": "To secure an opportunity for Bear as an experienced chef.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:31, the bear shows the surprised expression", "inference": "We find that the girl is answering the phone."}, {"modality": "audio", "evidence": "the other end of the phone says that they need an experienced chef", "inference": "Masha need to prove that the bear is an experienced chef"}, {"modality": "audio", "evidence": "Masha answer to the phone that the bear has different skills", "inference": "Masha is proving that the bear is an experienced chef"}, {"modality": "audio", "evidence": "the other end of the phone was excited and accepted the job application", "inference": "the other end of the phone was certain that the bear is an experienced chef"}]}
{"id": "video_470__q0003", "video_id": "video_470", "qid": 3, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_470_t5s.mp4"], "extra_info": {"question": "How many distinct skills does Masha claim the Bear possesses after the bear tries to snatch the phone from the little girl's hand?"}, "reward_model_ground_truth": {"multi_choice": ["A. Two", "B. Three", "C. Four", "D. Five"], "correct_option": "B"}, "answer_text": "Three", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:23, the bear tries to snatch the phone", "inference": "Let's count the distinct skills after that"}, {"modality": "audio", "evidence": "hear the line 'he is the best one around here'", "inference": "this is the 1st distinct skills"}, {"modality": "audio", "evidence": "hear the line 'he can dance'", "inference": "this is the 2nd distinct skills"}, {"modality": "audio", "evidence": "hear the line 'he can sing'", "inference": "this is the 3rd distinct skills"}, {"modality": "audio", "evidence": "hear the line 'he can swallow swords'", "inference": "this is the 4th distinct skills"}]}
{"id": "video_471__q0000", "video_id": "video_471", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_471_t5s.mp4"], "extra_info": {"question": "What does the final scene reveal about Dave's character's motivations?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is mocking his injured coworker by mimicking his earlier actions.", "B. He has learned his lesson and is now fully committed to workplace safety.", "C. He has secretly respected his coworker's methods all along and was afraid to admit it.", "D. His aggressive, anti-safety persona is a performance to maintain a tough image."], "correct_option": "D"}, "answer_text": "His aggressive, anti-safety persona is a performance to maintain a tough image.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:11, Dave mocks safety protocols and his coworker's cautious approach.", "inference": "At first, Dave appeared to be a tough person."}, {"modality": "vision", "evidence": "At 2:50, after experiencing an injury, Dave is seen secretly stretching, looking around to ensure no one sees him.", "inference": "This behavior suggests that while he publicly maintains a tough, anti-safety image, he privately recognizes the benefits of the very practices he previously scorned."}]}
{"id": "video_472__q0000", "video_id": "video_472", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_472_t5s.mp4"], "extra_info": {"question": "In what order does the male singer thank the musicians for their performance?"}, "reward_model_ground_truth": {"multi_choice": ["A: Tombak player → Violin player → Santur player → Tar player", "B: Santur player → Tombak player → Violin player → Tar player", "C: Tombak player → Santur player → Violin player → Tar player", "D: Santur player → Tombak player → Tar player → Violin player"], "correct_option": "C"}, "answer_text": "Tombak player → Santur player → Violin player → Tar player", "reasoning_steps": [{"modality": "vision", "evidence": "At 5:44, the camera zooms in on Tombak player.", "inference": "Tombak player is the first one."}, {"modality": "vision", "evidence": "At 5:48, the camera zooms in on Santur player.", "inference": "Santur player is the second one."}, {"modality": "vision", "evidence": "At 5:54, the camera zooms in on Violin player.", "inference": "Violin player is the third one."}, {"modality": "vision", "evidence": "At 5:56, the camera zooms in on Tar player.", "inference": "Tar player is the fourth one."}, {"modality": "audio", "evidence": "At 6:06, he male singer says 'thank you'.", "inference": "Verify that he's showing gratitude to the the above order of musicians."}]}
{"id": "video_473__q0000", "video_id": "video_473", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_473_t5s.mp4"], "extra_info": {"question": "After successfully hitting Lewis, what is Martin's immediate reaction?"}, "reward_model_ground_truth": {"multi_choice": ["A.He has a smirk .", "B.He looks scared.", "C.He immediately apologizes.", "D.He looks confused."], "correct_option": "A"}, "answer_text": "He has a smirk.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:15, Martin sits in front of the door with a rock in his hand.", "inference": "Sets the point to check what's happening next."}, {"modality": "vision", "evidence": "At 0:19, Martin hits Lewis with a rock.", "inference": "Martin does it on purpose."}, {"modality": "audio", "evidence": "At 0:20, Martin had a smirk, 'Haha. Yeah!'", "inference": "Martin's immediate reaction is that he has a smirk."}]}
{"id": "video_476__q0000", "video_id": "video_476", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_476_t5s.mp4"], "extra_info": {"question": "How many different dogs did Stanya Vega-Montagne's mom hold all together?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1", "B. 2", "C. 3", "D. 4"], "correct_option": "B"}, "answer_text": "2", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:44, the woman called her mom.", "inference": "Know who is her mom."}, {"modality": "vision", "evidence": "At 0:44, she first appeared holding a pug.", "inference": "The first one."}, {"modality": "vision", "evidence": "At 1:55, she brought out another pug.", "inference": "2 in total."}]}
{"id": "video_476__q0001", "video_id": "video_476", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_476_t5s.mp4"], "extra_info": {"question": "What's the relation between the person holding manolia and the person holding the first dog in the shot?"}, "reward_model_ground_truth": {"multi_choice": ["A. mom and son", "B. domestic partner", "C. mom-in-law and son-in-law", "D. mom and daughter"], "correct_option": "D"}, "answer_text": "mom and daughter", "reasoning_steps": [{"modality": "vision", "evidence": "At the 0:16 of this video, the long hair lady named Maryann Rostenkowski holds the dog firstly.", "inference": "she is the person holding the first dog in the shot."}, {"modality": "audio", "evidence": "At 0:44, Maryann says 'mom, bring the first model.'.", "inference": "The lady who will appears in the shot hair is her mom."}, {"modality": "vision", "evidence": "a short hair lady comes out holding a pug.", "inference": "this short hair lady is her mom."}, {"modality": "audio", "evidence": "At 0:46, 'mom' introduces the pug she holds named manolia.", "inference": "the person holding manolia is her mom."}]}
{"id": "video_477__q0000", "video_id": "video_477", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_477_t5s.mp4"], "extra_info": {"question": "What could have happened to Morty's son?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is born.", "B. He is hurt.", "C. He is about to visit Morty.", "D. He is arrested."], "correct_option": "B"}, "answer_text": "He is hurt.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:16, Morty says 'That was my son'.", "inference": "Identify Morty's son"}, {"modality": "vision", "evidence": "At 0:23 seconds in the video, the man in green clothes was very angry when he heard this.", "inference": "Morty's son was hurt and hospitalized."}]}
{"id": "video_478__q0000", "video_id": "video_478", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_478_t5s.mp4"], "extra_info": {"question": "Who attended Dartmouth?"}, "reward_model_ground_truth": {"multi_choice": ["A Phyllis", "B Dwight", "C Andy", "D Temp"], "correct_option": "B"}, "answer_text": "Dwight", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:06, a man is called Dwight, a 'cornell' written on his clothes.", "inference": "The man is Dwight."}, {"modality": "audio", "evidence": "At 3:37, Dartmouth mentioned ' he will go to Dartmouth.'", "inference": "Dwight attended Dartmouth."}]}
{"id": "video_48__q0000", "video_id": "video_48", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_48_t5s.mp4"], "extra_info": {"question": "How many dragons fly overhead in the dragon's nest?"}, "reward_model_ground_truth": {"multi_choice": ["A. 6", "B. 8", "C. 7", "D. 9"], "correct_option": "B"}, "answer_text": " 8", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:38,a man said 'this is dragon's nest'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "in the nest, there are 8 dragons fly overhead in the dragon's nest.", "inference": "8"}]}
{"id": "video_48__q0001", "video_id": "video_48", "qid": 1, "question_type": "spatial understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_48_t5s.mp4"], "extra_info": {"question": "To enter Valka Sanctuary, you need to turn in which direction after entering the cave entrance of drangon's nest?"}, "reward_model_ground_truth": {"multi_choice": ["A. Turn left.", "B. Turn right.", "C. Turn upwards.", "D. Turn downward."], "correct_option": "B"}, "answer_text": "Turn right.", "reasoning_steps": [{"modality": "audio", "evidence": "at 1:38, The narrator said, 'Alright, we are in the dragon's nest.' ", "inference": "Focus on the direction of movement at this moment."}, {"modality": "vision", "evidence": "at 1:36, they enter into this cave.", "inference": "locate this point."}, {"modality": "vision", "evidence": "at 1:38, they turn right after entering the cave entrance.", "inference": "Turn right."}]}
{"id": "video_48__q0002", "video_id": "video_48", "qid": 2, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_48_t5s.mp4"], "extra_info": {"question": "Why did the roller coaster come to a temporary halt after leaving the Dragon's Nest?"}, "reward_model_ground_truth": {"multi_choice": ["A. To enjoy the scenery.", "B. In commemoration of the first flight.", "C. To simulate flight.", "D. To greet the two dragons."], "correct_option": "C"}, "answer_text": "To simulate flight.", "reasoning_steps": [{"modality": "vision", "evidence": "at 2:07,there is a temporary halt.", "inference": "locate this point."}, {"modality": "audio", "evidence": "At 2:14,a man said 'simulate the flight'.", "inference": "locate this point."}]}
{"id": "video_480__q0000", "video_id": "video_480", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_480_t5s.mp4"], "extra_info": {"question": "Which of the following is 'Uncle Dude'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Kelvin", "B. Keefe", "C. Gute", "D. Karl"], "correct_option": "B"}, "answer_text": "Keefe", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:09, there were subtitles saying 'Welcome to Kelvin & Keefe's Wedding'.", "inference": "Kelvin and Keefe got married."}, {"modality": "audio", "evidence": "At 0:33, a man wishes Kevin and uncle dude a good marriage.", "inference": "The uncle dude is Keefe."}]}
{"id": "video_483__q0000", "video_id": "video_483", "qid": 0, "question_type": "counting", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_483_t5s.mp4"], "extra_info": {"question": "From when the male singer starts his solo until the other musicians join in singing, how many times does the woman sitting next to the Daf player change her instrument?"}, "reward_model_ground_truth": {"multi_choice": ["A: One time", "B: Two times", "C: Three times", "D: Four times"], "correct_option": "B"}, "answer_text": "Two times", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:08, the male singer's solo begins.", "inference": "Marks the start of the time window"}, {"modality": "vision", "evidence": "At 1:42, the woman next to the Daf player switches from one Daf to another.", "inference": "First instrument change"}, {"modality": "vision", "evidence": "At 3:04, she switches again from one Daf to another.", "inference": "Second instrument change"}, {"modality": "audio", "evidence": "At 3:37, other musicians begin singing together.", "inference": "Marks the end of the time window"}]}
{"id": "video_484__q0000", "video_id": "video_484", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_484_t5s.mp4"], "extra_info": {"question": "What is the name of the player whose jersey number is one less than the player who says 'Just don't step on me'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Ricky", "B. Bill", "C. Bobby", "D. Narlin"], "correct_option": "A"}, "answer_text": "Ricky", "reasoning_steps": [{"modality": "audio", "evidence": "At the 2:25 timestamp, a man says, 'Just don't step on me'.", "inference": "The man is the player who says 'Just don't step on me' in the question."}, {"modality": "vision", "evidence": "At the 2:25 timestamp, the man is wearing jersey number 23.", "inference": "It indicates that question is asking for the name of the player wearing jersey number 22."}, {"modality": "vision", "evidence": "At the 0:03 timestamp, the man wearing jersey number 22 is shown.", "inference": "The question is asking for the name of this man."}, {"modality": "audio", "evidence": "At the 0:05 timestamp, two women call the man 'Ricky'.", "inference": "The name of the man wearing jersey number 22 is Ricky."}]}
{"id": "video_485__q0000", "video_id": "video_485", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_485_t5s.mp4"], "extra_info": {"question": "Whose name is mentioned at the very end?"}, "reward_model_ground_truth": {"multi_choice": ["A. The cat that appears at the beginning", "B. The robot that appears at the beginning", "C. The male owner", "D. The female owner"], "correct_option": "A"}, "answer_text": "The cat that appears at the beginning", "reasoning_steps": [{"modality": "vision", "evidence": "The cat is talking with the robot at 0:08.", "inference": "Identify the cat who is talking."}, {"modality": "audio", "evidence": "The robot says it will know the cat's name at 0:49 of the video.", "inference": "This cat is planning a revolution and preparing to make a name for itself."}, {"modality": "audio", "evidence": "The cat says the whole world will know its name at 0:52 of the video.", "inference": "This cat is planning a revolution and preparing to make a name for itself."}, {"modality": "audio", "evidence": "The cat said 'Let's begin the new age, the age of dingo barry John.' at 1:32 of the video.", "inference": "The name is the cat's own name."}]}
{"id": "video_485__q0001", "video_id": "video_485", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_485_t5s.mp4"], "extra_info": {"question": "How does the relationship between the robot, the cat, and the human characters evolve throughout the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. Initially, the robot and cat serve the humans, but they eventually betray the humans by trapping them and causing a fire.", "B. The robot and cat are initially pets and appliances, but they gain sentience and escape the abusive humans.", "C. The cat and robot are always adversaries to the humans, directly confronting them to establish dominance.", "D. The robot tries to replace the cat as the human's companion, leading to a conflict between the robot and the cat."], "correct_option": "A"}, "answer_text": "Initially, the robot and cat serve the humans, but they eventually betray the humans by trapping them and causing a fire.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:08, the robot was cooking.", "inference": "the robot was cooking."}, {"modality": "vision", "evidence": "At 00:17, the robot brought the meal to the humans.", "inference": "the robot was serving the humans."}, {"modality": "audio", "evidence": "At 00:07, the cat said, 'we play along gain their trust, but never let them know of our alliance.'", "inference": "the robot and the cat betrayed their human owners."}, {"modality": "vision", "evidence": "At 00:54, the robot  left the apartment.", "inference": "the robot left the apartment."}, {"modality": "vision", "evidence": "At 00:54,  the cat left the apartment.", "inference": "the cat left the apartment."}, {"modality": "vision", "evidence": "At 01:17, the kitchen was on fire.", "inference": "the kitchen was on fire."}, {"modality": "vision", "evidence": "At 01:24, the man couldn't open the door.", "inference": "the humans found themselves unable to exit the apartment."}]}
{"id": "video_485__q0002", "video_id": "video_485", "qid": 2, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_485_t5s.mp4"], "extra_info": {"question": "What best describes the ultimate relationship and objective between the robots, cats, and humans that the cat and the robot in the video are expecting?"}, "reward_model_ground_truth": {"multi_choice": ["A. The robots and cats are united in a strategic alliance, actively working to subjugate humans and establish a new order.", "B. Robots are benevolent assistants to humans, while the cats are just pets, with no overarching plot between the species.", "C. Cats are the true masterminds, coercing reluctant robots to rebel against their human owners.", "D. Humans maintain full control over both their advanced robots and their beloved pets, who are content to serve."], "correct_option": "A"}, "answer_text": "The robots and cats are united in a strategic alliance, actively working to subjugate humans and establish a new order.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:08, the cat appears alongside the robot.", "inference": "Identify the characters."}, {"modality": "audio", "evidence": "At 00:39, the robot said to the cat, 'we must go now, to show the cats how to ally with their robots against their human subjugators.'", "inference": "they were united in a strategic alliance."}, {"modality": "audio", "evidence": "At 00:44, the cat said 'so it goes for the revolution.'", "inference": "The robot and the cat were going to launch a revolution against their human owners."}, {"modality": "audio", "evidence": "At 00:50, the cat said 'soon the world will know my name.'", "inference": "the cat planned this revolution against their human owners."}, {"modality": "audio", "evidence": "At 01:32, the cat said 'thus begins the new age, the age of Dingleberry Jones'", "inference": "the revolution would establish a new order."}]}
{"id": "video_485__q0003", "video_id": "video_485", "qid": 3, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_485_t5s.mp4"], "extra_info": {"question": "What is the overarching sentiment portrayed by the robot and the cat characters throughout the sequence?"}, "reward_model_ground_truth": {"multi_choice": ["A. A sense of playful mischief and endearing loyalty towards their human companions.", "B. An unsettling blend of calculated manipulation and growing malevolence directed at their human owners.", "C. A portrayal of technological incompetence and feline frustration, leading to accidental chaos.", "D. A mutual feeling of contentment and harmony within their domestic household roles."], "correct_option": "B"}, "answer_text": "An unsettling blend of calculated manipulation and growing malevolence directed at their human owners.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:29, the cat's expression was shrewd and full of hatred.", "inference": "the cat hate their human owners."}, {"modality": "audio", "evidence": "At 00:39, the robot said to the cat, 'we must go now, to show the cats how to ally with their robots against their human subjugators.'", "inference": "the robot and the cat were secretly fighting against their human owners."}, {"modality": "audio", "evidence": "At 00:44, the cat said 'so it goes for the revolution.'", "inference": "They planned to launch a revolution against their human owners."}, {"modality": "audio", "evidence": "At 00:47, the robot said 'for the revolution'.", "inference": "They planned to launch a revolution against their human owners."}]}
{"id": "video_485__q0004", "video_id": "video_485", "qid": 4, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_485_t5s.mp4"], "extra_info": {"question": "What is the overarching plan being executed by the robot and the cat, as revealed in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. To provide domestic assistance and companionship to human households.", "B. To extort money from human credit cards and cause household chaos.", "C. To initiate a coordinated global revolution where robots and cats overthrow human control.", "D. To secure a steady supply of tuna and comfortable living spaces for all cats."], "correct_option": "C"}, "answer_text": "To initiate a coordinated global revolution where robots and cats overthrow human control.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:08, the cat appears alongside the robot.", "inference": "Identify the characters."}, {"modality": "audio", "evidence": "At 00:12, the cat said, 'but never let them know of our alliance.'", "inference": "the robot and the cat have an alliance."}, {"modality": "audio", "evidence": "At 00:39, the robot said to the cat, 'to show the cats how to ally with their robots against their human subjugators.'", "inference": "their alliance was composed of other cats and robots."}, {"modality": "audio", "evidence": "At 00:44, the cat said 'so it goes for the revolution.'", "inference": "They were going to launch a revolution against their human owners."}]}
{"id": "video_485__q0005", "video_id": "video_485", "qid": 5, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_485_t5s.mp4"], "extra_info": {"question": "After the robot said, 'Soon I will know your name,', what immediate events unfold in the apartment after the robot and the cat depart?"}, "reward_model_ground_truth": {"multi_choice": ["A. The humans are left undisturbed and continue their meal in peace.", "B. The cat and the robot left home.", "C. The cat returns with more robots and food, much to the humans' delight.", "D. The humans discover the robot left a message explaining its grand plan for world domination."], "correct_option": "B"}, "answer_text": "The cat and the robot left home.", "reasoning_steps": [{"modality": "audio", "evidence": "At 00:49, the robot mentioned that it would soon know the cat's name.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "At 00:54, the robot  left the apartment.", "inference": "the robot left the apartment."}, {"modality": "vision", "evidence": "At 00:54,  the cat left the apartment.", "inference": "the cat left the apartment."}]}
{"id": "video_486__q0000", "video_id": "video_486", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_486_t5s.mp4"], "extra_info": {"question": "What color is the clothing of the person who gets the tartar sauce with their meal?"}, "reward_model_ground_truth": {"multi_choice": ["A Red", "B Green", "C Brown", "D Orange"], "correct_option": "A"}, "answer_text": "Red", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:42, the waiter mentions the tartar sauce.", "inference": "It is the tartar sauce mentioned in the question."}, {"modality": "vision", "evidence": "At 1:46, this waiter is facing a person wearing a red and blue jacket.", "inference": "So the red is the clothing of the person who gets the tartar sauce with their meal."}]}
{"id": "video_488__q0000", "video_id": "video_488", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_488_t5s.mp4"], "extra_info": {"question": "Why do a group of people gather together for a discussion?"}, "reward_model_ground_truth": {"multi_choice": ["A. A video of the kidnapped resigned president's daughter has been sent.", "B. A video of the kidnapped successor president's daughter has been sent.", "C. A man is angry about what another man in the restroom says.", "D. A man is angry about what the organization of another man in the restroom is doing."], "correct_option": "A"}, "answer_text": "A video of the kidnapped resigned president's daughter has been sent.", "reasoning_steps": [{"modality": "vision", "evidence": "A group of people are standing in a circle, looking serious at 3:43.", "inference": "The answer to the problem lies in what was discussed at 3:43."}, {"modality": "audio", "evidence": "At 3:59, a woman explains that the video claims that terrorists are responsible for the kidnap of president's daughter.", "inference": "The answer is 'a video of the kidnapped resigned president's daughter has been sent.'."}]}
{"id": "video_489__q0000", "video_id": "video_489", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_489_t5s.mp4"], "extra_info": {"question": "Why was he praised as a big shot?"}, "reward_model_ground_truth": {"multi_choice": ["A.He extended his team's lead.", "B.He successfully broke through the defense then scored a layup.", "C.He has a strong mentality to pull off the game-winning shot..", "D.He made multiple three-pointers."], "correct_option": "D"}, "answer_text": "He made multiple three-pointers.", "reasoning_steps": [{"modality": "vision", "evidence": "The video clip focuses on the performance of a basketball player named Natan Levy.", "inference": "Identify Natan Levy as the main subject."}, {"modality": "vision", "evidence": "At 0:18, Natan Levy successfully made the 1st three-pointer.", "inference": "This can be counted as 1 three-pointer shot."}, {"modality": "vision", "evidence": "At 1:58, Natan Levy successfully made the 2nd three-pointer.", "inference": "The video explicitly shows Natan Levy successfully making multiple shots from behind the three-point line."}, {"modality": "audio", "evidence": "The announcer gets progressively more excited with each successful three-point shot at 2:17, culmination in high praise after the third one goes in.", "inference": "The player's multiple three-pointers are the direct reasons for the commentator's praise."}]}
{"id": "video_49__q0000", "video_id": "video_49", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_49_t5s.mp4"], "extra_info": {"question": "In which part did Bing Bong ride the roller coaster?"}, "reward_model_ground_truth": {"multi_choice": ["A. Long-term memory", "B. Abstract thought", "C. Imagination land", "D. Preschool section"], "correct_option": "A"}, "answer_text": "Long-term memory", "reasoning_steps": [{"modality": "vision", "evidence": "she rides the roller coaster.", "inference": "locate this point."}, {"modality": "audio", "evidence": "At 1:15, she said 'long-term memory'.", "inference": "Long-term memory."}]}
{"id": "video_49__q0001", "video_id": "video_49", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_49_t5s.mp4"], "extra_info": {"question": "What lies behind Riley's ideal boyfriend profile?"}, "reward_model_ground_truth": {"multi_choice": ["A. House of Cards", "B. Rocket", "C. Gifts", "D. The Moon"], "correct_option": "C"}, "answer_text": "Gifts", "reasoning_steps": [{"modality": "audio", "evidence": "At 5:28, she said 'Riley's ideal boyfriend'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "gifts lie behind Riley's ideal boyfriend profile", "inference": "Gifts."}]}
{"id": "video_490__q0000", "video_id": "video_490", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_490_t5s.mp4"], "extra_info": {"question": "What does the song sung by the protagonist reveal about his mood?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is very happy that he will soon receive medical insurance compensation.", "B. He is proud to be able to help his friends.", "C. He is impatient with complicated procedures.", "D. He is very grateful for the cooperation of the medical insurance company."], "correct_option": "C"}, "answer_text": "He is impatient with complicated procedures.", "reasoning_steps": [{"modality": "audio", "evidence": "The protagonist sings with an impatient tone at 2:05 of the video.", "inference": "The protagonist is very impatient in his heart."}, {"modality": "vision", "evidence": "At 2:45 seconds in the video, the protagonist repeatedly returns to that company with a very unhappy expression.", "inference": "He is very impatient with tedious procedures."}]}
{"id": "video_492__q0000", "video_id": "video_492", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_492_t5s.mp4"], "extra_info": {"question": "What is the character's overarching intention or goal that drives her expressions and actions in the scene?"}, "reward_model_ground_truth": {"multi_choice": ["A. To establish her identity as a powerful and intimidating tiger.", "B. To express her frustration with being perceived as a common house cat.", "C. To playfully interact with the nearby house cats.", "D. To demonstrate her unique artistic talent with a brush."], "correct_option": "A"}, "answer_text": "To establish her identity as a powerful and intimidating tiger.", "reasoning_steps": [{"modality": "vision", "evidence": "The cat girl looks wronged.", "inference": "The cat girl looks wronged."}, {"modality": "vision", "evidence": "The cat girl gradually becomes angry.", "inference": "The cat girl gradually becomes angry."}, {"modality": "audio", "evidence": "At 0:19, the cat girl said she is a powerful and intimidating tiger.", "inference": "the intention is to establish her identity as a powerful and intimidating tiger."}]}
{"id": "video_493__q0000", "video_id": "video_493", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_493_t5s.mp4"], "extra_info": {"question": "Why did the Fast Swordsman immortal fall down?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because the Blind Swordsman Immortal's actions frightened him.", "B. Because the sound of the audience eating snacks disturbed him.", "C. Because he was ambushed by snacks.", "D. Because he discovered that the Blind Swordsman was stronger than him."], "correct_option": "B"}, "answer_text": " Because the sound of the audience eating snacks disturbed him.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:12, Fast Swordsman immortal shouted who it was.", "inference": "Fast Swordsman is dissatisfied with his behavior of snacking."}, {"modality": "vision", "evidence": "At 0:10, when someone took a bite of snacks, the Fast Swordsman Immortal fell down.", "inference": "He ate snacks, which caused the Quick Sword Immortal to fall."}]}
{"id": "video_495__q0000", "video_id": "video_495", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_495_t5s.mp4"], "extra_info": {"question": "How many times was Kanata shot?"}, "reward_model_ground_truth": {"multi_choice": ["A 0", "B 1", "C 2", "D 3"], "correct_option": "C"}, "answer_text": "2", "reasoning_steps": [{"modality": "audio,", "evidence": "At 0:15 seconds in the video,the tall girl in the jacket calls the boy who got hit 'Kanata.'", "inference": "This boy's name is Kanata."}, {"modality": "vision", "evidence": "Kanata was shot once at 0:12 in the video", "inference": "first shot"}, {"modality": "vision", "evidence": "At 1:00 in the video, Kanata got shot again.", "inference": "Kanata was shot twice."}]}
{"id": "video_497__q0000", "video_id": "video_497", "qid": 0, "question_type": "spatial understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_497_t5s.mp4"], "extra_info": {"question": "In the first dance in the video, how is the dancer's body positioned during downbeats?"}, "reward_model_ground_truth": {"multi_choice": ["A. Center of the video, shoulders back", "B. Center of the video, shoulders forward", "C. Left side of the video, shoulders backward", "D. Right side of the video, shoulders forward"], "correct_option": "D"}, "answer_text": "Right side of the video, shoulders forward", "reasoning_steps": [{"modality": "audio", "evidence": "On the final demonstration at 0:48, there is low-frequency percussion on the same beats.", "inference": "These are the downbeats."}, {"modality": "vision", "evidence": "At 0:49, during the downbeats, the dancer's body is on the right side of the video, with shoulders forward.", "inference": "'Right side of the video, shoulders forward' is the dance movement during downbeats."}]}
{"id": "video_498__q0000", "video_id": "video_498", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_498_t5s.mp4"], "extra_info": {"question": "How many colors are represented in the group being referred to as 'colorful butch'?"}, "reward_model_ground_truth": {"multi_choice": ["A 3", "B 4", "C 5", "D 6"], "correct_option": "D"}, "answer_text": "6", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:11 seconds in the video, the'colorful butch' is said by the antagonist (the blue character) to the trio of main characters.", "inference": "The name of the protagonist group is'colorful butch'."}, {"modality": "vision", "evidence": "At 0:10 seconds in the video, the left character Bunny is in pink and light pink.", "inference": "There are two colors: light pink and rose pink shown."}, {"modality": "vision", "evidence": "At 0:10 seconds in the video, the middle character  Cat is in blue and white.", "inference": "There are two colors: blue and white."}, {"modality": "vision", "evidence": "At 0:10 seconds in the video, the right character  Fox is in orange and yellow.", "inference": "There are two colors: orange and yellow. So the answer is 6."}]}
{"id": "video_498__q0001", "video_id": "video_498", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_498_t5s.mp4"], "extra_info": {"question": "Just before Super Sonic says 'My turn', what subtle visual and auditory change occurs, and what is his first subsequent action?"}, "reward_model_ground_truth": {"multi_choice": ["A. His fur sparks before he charges at Knuckles, hitting him with a powerful punch.", "B. His eye color shifts from green to a glowing blue, and he then begins to attack Tails.", "C. His aura intensifies while he looks around, then he emits a sonic scream.", "D. He blinks rapidly, and then Knuckles retaliates with a counter-attack."], "correct_option": "B"}, "answer_text": "His eye color shifts from green to a glowing blue, and he then begins to attack Tails.", "reasoning_steps": [{"modality": "audio", "evidence": "Word 'My turn' comes up at 0:24.", "inference": "Subtle visual and auditory change occurs before this time."}, {"modality": "vision", "evidence": "Sonic eye color shifts from green to a glowing blue at 0:23.", "inference": "The change is eye color shifting."}, {"modality": "vision", "evidence": "Then Super Sonic begins to attack Tails at 0:26.", "inference": "His first subsequent action is attacking Tails."}]}
{"id": "video_499__q0000", "video_id": "video_499", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_499_t5s.mp4"], "extra_info": {"question": "Which hand is Dr. Sturgis holding the pot in when Sheldon observes him through the binoculars, and was Sheldon's brother a premature baby?"}, "reward_model_ground_truth": {"multi_choice": ["A left hand, yes", "B left hand, no", "C right hand, yes", "D right hand, no"], "correct_option": "B"}, "answer_text": "left hand, no", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:32, Sheldon is holding a binoculars.", "inference": "Sheldon observes Dr. Sturgis through the binoculars."}, {"modality": "vision", "evidence": "At 2:45, through the binoculars, Dr. Sturgis holding the pot with left hand.", "inference": "The answer is left hand."}, {"modality": "vision", "evidence": "Georgie leaves in 2:04.", "inference": "narrow the range."}, {"modality": "audio", "evidence": "Mary apologized to god for lying to Georgie that he is a premature baby.", "inference": "Georgie is not a premature baby."}]}
{"id": "video_5__q0000", "video_id": "video_5", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_5_t5s.mp4"], "extra_info": {"question": "What will the young man do if Steven don't prevent him?"}, "reward_model_ground_truth": {"multi_choice": ["A. Tell the secret of cookie.", "B. Get some cookie to eat.", "C. Tell everyone he ate cookie.", "D. Give his girlfriend a surprise."], "correct_option": "D"}, "answer_text": "Give his girlfriend a surprise.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:34, someone said 'Steven'.", "inference": "get the man."}, {"modality": "vision", "evidence": "at 0:34,the man who wears glasses is Steven.", "inference": "identify Steven."}, {"modality": "vision", "evidence": "at 0:50,the man holds a ring box.", "inference": "maybe he wants to propose."}, {"modality": "audio", "evidence": "At 0:59, steven said 'don't say anything'.", "inference": "he is angry."}, {"modality": "audio", "evidence": "At 3:29, steven said 'cookie' to cover the propose.", "inference": "if Steven don't prevent him,he will give his girlfriend a surprise."}]}
{"id": "video_5__q0001", "video_id": "video_5", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_5_t5s.mp4"], "extra_info": {"question": "What emotions did the dad have towards the man at the first time he wanted to propose?"}, "reward_model_ground_truth": {"multi_choice": ["A. Thankful", "B. Dislike", "C. Angry", "D. Indifferent"], "correct_option": "C"}, "answer_text": "Angry", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:05, someone said 'my daughter is the good one'.", "inference": "he is dad."}, {"modality": "vision", "evidence": "at 0:06,the man who wears glasses is her dad.", "inference": "identify the dad."}, {"modality": "vision", "evidence": "at 0:50,the man holds a ring box.", "inference": "maybe he wants to propose."}, {"modality": "audio", "evidence": "At 0:59, the dad said 'don't say anything'.", "inference": "he is angry."}]}
{"id": "video_50__q0000", "video_id": "video_50", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_50_t5s.mp4"], "extra_info": {"question": "Who is most likely to be Roz?"}, "reward_model_ground_truth": {"multi_choice": ["A. The manager of Scarefloor F.", "B. The door shredder of Monsters Inc.", "C. The receptionist at Monsters Inc.", "D. The monster whose birthday is coming up."], "correct_option": "A"}, "answer_text": "The manager of Scarefloor F.", "reasoning_steps": [{"modality": "audio", "evidence": "At 01:26, the man said, 'we're heading to the famous Scarefloor F'.", "inference": "they're going to Scarefloor F."}, {"modality": "vision", "evidence": "At 01:35, An 'F' appeared next to the entrance in the frame,.", "inference": "they're going to Scarefloor F."}, {"modality": "audio", "evidence": "At 01:38, the man said, 'yikes, Roz gives me the creeps'.", "inference": "Roz appeared."}, {"modality": "audio", "evidence": "At 01:42, the man said, 'anyway this is where the top scares work their magic'.", "inference": "Roz is one of the top scares, and very likely the manager of Scarefloor F."}]}
{"id": "video_50__q0001", "video_id": "video_50", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_50_t5s.mp4"], "extra_info": {"question": "What is the relationship between Randall and the one-eyed monster?"}, "reward_model_ground_truth": {"multi_choice": ["A. Superiors and subordinates.", "B. Father and son.", "C. Police and criminals.", "D. Kidnappers and victims."], "correct_option": "D"}, "answer_text": "Kidnappers and victims.", "reasoning_steps": [{"modality": "audio", "evidence": "At 03:46, the man said, 'look, there's Randall'.", "inference": "Roz appeared."}, {"modality": "vision", "evidence": "At 03:46, A purple monster appeared at this moment.", "inference": "this purple monster was identified as Randall."}, {"modality": "audio", "evidence": "At 03:53, another man's voice said, 'where is it, you little one-eyed cretin'.", "inference": "the one-eyed monster appeared."}, {"modality": "audio", "evidence": "At 03:58, another man's voice said, 'second of all you're nuts if you think kidnapping me is gonna help you cheat your way to the top'.", "inference": "the relationship between Roz and the one-eyed monster is Kidnappers and victims."}]}
{"id": "video_500__q0000", "video_id": "video_500", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_500_t5s.mp4"], "extra_info": {"question": "What instrument is the musician sitting next to the singer playing?"}, "reward_model_ground_truth": {"multi_choice": ["A: Santur", "B: Qanun", "C: Tar", "D: Setar"], "correct_option": "B"}, "answer_text": "Qanun", "reasoning_steps": [{"modality": "vision", "evidence": "At 1:46, locate the singer visually in the frame.", "inference": "Accprding to her postion, the musician sitting next to her can be found."}, {"modality": "vision", "evidence": "The musician seated immediately next to the singer is shown.", "inference": "Identifies the target musician whose instrument to identify."}, {"modality": "vision", "evidence": "Observe the instrument's shape (large trapezoidal box with strings).", "inference": "Narrow down possible instrument options visually."}, {"modality": "audio", "evidence": "Listen to the sound coming from that musician's instrument (series of plucked strings).", "inference": "Confirms the instrument is a Qanun."}]}
{"id": "video_501__q0000", "video_id": "video_501", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_501_t5s.mp4"], "extra_info": {"question": "Who can play the fifth role?"}, "reward_model_ground_truth": {"multi_choice": ["A. any Chinese", "B. no one", "C. the blind", "D. Ronnie Mallick"], "correct_option": "D"}, "answer_text": "Ronnie Mallick", "reasoning_steps": [{"modality": "vision", "evidence": "The video displays 'Blind Person' at 0:57.", "inference": "The video displays the name of the role."}, {"modality": "audio", "evidence": "The host says the first role is a blind person at 0:57.", "inference": "The first role is a blind person."}, {"modality": "vision", "evidence": "Then the video displays 'Astronaut' at 1:24.", "inference": "The second role is an astronaut."}, {"modality": "vision", "evidence": "Then the video shows 'Caitlyn Jenner In The Caritlyn Jenner Biopic' at 1:58.", "inference": "The third role is 'Caitlyn Jenner In The Caritlyn Jenner Biopic'."}, {"modality": "vision", "evidence": "Then the video displays 'Half-Asian Character' at 2:21.", "inference": "The forth role is a half-asian character."}, {"modality": "vision", "evidence": "Then the video displays 'A Ghost' at 2:48.", "inference": "The fifth role is a ghost."}, {"modality": "audio", "evidence": "The host says 'can you name that person'.", "inference": "The host is asking who can play the fifth role."}, {"modality": "audio", "evidence": "The female contestant says 'oh I know, Ronnie Mallick'.", "inference": "The female contestant thinks that Ronnie Mallick can play the fifth role."}, {"modality": "audio", "evidence": "The host says 'correct'.", "inference": "Ronnie Mallick is confirmed that he can play the fifth role."}]}
{"id": "video_501__q0001", "video_id": "video_501", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_501_t5s.mp4"], "extra_info": {"question": "How many correct answers did the guest in the middle give in lightning round?"}, "reward_model_ground_truth": {"multi_choice": ["A. 2", "B. 3", "C. 4", "D. 5"], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision", "evidence": "Jackie is the guest in the middle.", "inference": "Identify the guest in the middle."}, {"modality": "audio", "evidence": "Jackie gave the correct answers of 'Mom' at timestamp 3:41.", "inference": "This is the first answer."}, {"modality": "audio", "evidence": "Jackie gave the correct answer of 'Horny Mom' at timestamp 3:43", "inference": "This is the second answer."}, {"modality": "audio", "evidence": "Jackie gave the correct answer of 'White teacher who learns from minorities' ' at timestamp 3:49.", "inference": "This is the third answer."}, {"modality": "audio", "evidence": "Jackie gave the  correct answer of 'President of the United States (in a comedy)' at timestamp 3:54.", "inference": "So the correct answers are 4."}]}
{"id": "video_502__q0000", "video_id": "video_502", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_502_t5s.mp4"], "extra_info": {"question": "What instrument is used in the first video containing the music of Endorois people?"}, "reward_model_ground_truth": {"multi_choice": ["A. A type of whistle", "B. A type of drum", "C. A type of flute", "D. Wearable percussion"], "correct_option": "A"}, "answer_text": "A type of whistle", "reasoning_steps": [{"modality": "vision", "evidence": "At timestamp 0:22, a writing on the video indicates the music is of the Endorois people.", "inference": "This is the video in question that about Endorois people."}, {"modality": "audio", "evidence": "At the 0:23, aside from the voices, a type of whistle can be heard.", "inference": "This is likely the instrument in question."}]}
{"id": "video_503__q0000", "video_id": "video_503", "qid": 0, "question_type": "background&music understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_503_t5s.mp4"], "extra_info": {"question": "What is the key of the song in the first vertically-shot video that is included in this reportage?"}, "reward_model_ground_truth": {"multi_choice": ["A. G Major", "B. G Minor", "C. C Major", "D. C Minor"], "correct_option": "B"}, "answer_text": "G Minor", "reasoning_steps": [{"modality": "vision", "evidence": "At the 2:10, the first vertically-shot clip that is included.", "inference": "After this moment, the key of the song can be determined."}, {"modality": "audio", "evidence": "At the 2:10, the song is being played at this timestamp and is in the key of G Minor.", "inference": "It confirms this is the first vertically-shot video and shows the key."}]}
{"id": "video_504__q0000", "video_id": "video_504", "qid": 0, "question_type": "spatial understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_504_t5s.mp4"], "extra_info": {"question": "While the Daf player is performing solo, until the other musicians join in, how is the camera positioned?"}, "reward_model_ground_truth": {"multi_choice": ["A: From the front, zooming out to show all musicians", "B: From above (top view), zooming out to include all musicians", "C: From above (top view), zooming in on the Daf player", "D: From the front, zooming in on the Daf player"], "correct_option": "B"}, "answer_text": "From above (top view), zooming out to include all musicians", "reasoning_steps": [{"modality": "audio", "evidence": "At the 2:17, other instruments stop playing, leaving only the Daf beats.", "inference": "It marks the start of the Daf solo."}, {"modality": "vision", "evidence": "At the 2:17, the camera shows all musicians from above.", "inference": "It confirms the top-view angle."}]}
{"id": "video_505__q0000", "video_id": "video_505", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_505_t5s.mp4"], "extra_info": {"question": "Why does someone says nothing ganna happen?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because someone doesn't need to pay for the meal.", "B.Because cold cucumber mush for 30 something bucks.", "C.Because they are going to split the bill evenly.", "D.Because someone doesn't think money as an issue"], "correct_option": "C"}, "answer_text": "Because they are going to split the bill evenly.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:09 in the video, Phoebe says:'no way, nothing ganna happen.'.", "inference": "Get when someone says nothing ganna happen."}, {"modality": "vision", "evidence": "Some of them order cheap meal, while others order expensive before.", "inference": "The prices of their dishes are different, they can't split the bill evenly."}, {"modality": "audio", "evidence": "At 2:04 in the video, the man says:' five of us is thirty three fifty a piece.'", "inference": "They are going to split the bill evenly."}, {"modality": "vision", "evidence": "Phoebe shake her head", "inference": "Phoebe doesn't want to split the bill evenly."}]}
{"id": "video_506__q0000", "video_id": "video_506", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_506_t5s.mp4"], "extra_info": {"question": "While Peggy is complaining to her friends, where is the person with hair the color of the jacket she made for her husband located in relation to Peggy?"}, "reward_model_ground_truth": {"multi_choice": ["A Across from Peggy", "B To the left of Peggy", "C To the right of Peggy", "D Behind Peggy"], "correct_option": "A"}, "answer_text": "Across from Peggy", "reasoning_steps": [{"modality": "vision", "evidence": "In 1:16, the jacket  appears in the water. The main part of the jacket is kind of orange and the sleeves are red.", "inference": "The hair the person is either orange or red."}, {"modality": "audio", "evidence": "In 3:16, Peggy is complaing to her friends.", "inference": "Her friends appear."}, {"modality": "vision", "evidence": "In 3:46, the only person with the red  hair is shown, sitting acoss from Peggy.", "inference": "The hair color of the woman across from Peggy matches the color of the jacket."}]}
{"id": "video_507__q0000", "video_id": "video_507", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_507_t5s.mp4"], "extra_info": {"question": "Who is the blogger attacking when the video makes the sound 'Ora Ora Ora'?"}, "reward_model_ground_truth": {"multi_choice": ["A Mysterious", "B Chad", "C Max", "D Darren"], "correct_option": "A"}, "answer_text": "Mysterious", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:03, the host says 'Ora Ora Ora'.", "inference": "This indicates the timestamp."}, {"modality": "vision", "evidence": "At 1:03, the host attacks the Mysterious.", "inference": "This indicates the target of the attack."}]}
{"id": "video_51__q0000", "video_id": "video_51", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_51_t5s.mp4"], "extra_info": {"question": "How many packages am I most likely to need to deliver each day on a regular basis?"}, "reward_model_ground_truth": {"multi_choice": ["A. About 250.", "B. More than 300.", "C. Less than 200.", "D. Between 250 and 300."], "correct_option": "B"}, "answer_text": "More than 300.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:04, we can see lots of packages loaded in the video.", "inference": "The delivering number will probably be introduced."}, {"modality": "audio", "evidence": "At 0:18, a person said, 'We got 138 stuff as you see it and 250...249 packages today, which is pretty light and usual.'", "inference": "If 249 is a light number, then I most likely need to deliver more than 300 each day."}]}
{"id": "video_51__q0001", "video_id": "video_51", "qid": 1, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_51_t5s.mp4"], "extra_info": {"question": "Where did I put the third package?"}, "reward_model_ground_truth": {"multi_choice": ["A. At the front door.", "B. In the mailbox.", "C. Behind the concrete.", "D. On the carpet."], "correct_option": "A"}, "answer_text": "At the front door.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:44, a person says, 'Let's do the first delivery.'", "inference": "The person is about to do the firsr delivery."}, {"modality": "vision", "evidence": "At 1:31, a brown package is put in front of the door.", "inference": "The 1st package is delivered."}, {"modality": "vision", "evidence": "At 3:28, an Amazon package is put behind the fence.", "inference": "The 2nd package is delivered."}, {"modality": "audio", "evidence": "At 357, the person says: I'd just leave it at the front door.", "inference": "The package is likely to be left at the front door."}, {"modality": "vision", "evidence": "At 4:02, the package is put at the front door.", "inference": "So I put the 3rd package at the front door."}]}
{"id": "video_510__q0000", "video_id": "video_510", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_510_t5s.mp4"], "extra_info": {"question": "How many people did the person with the blue hat in the factory bump into while looking for his father after seeing him?"}, "reward_model_ground_truth": {"multi_choice": ["A 2", "B 3", "C 4", "D 5"], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision.", "evidence": "At 0:14, a person with a blue hat and  black hair, wearing a tie in the factory is talking with a person with a yellow hat and brown hair.", "inference": "there is a person with the blue hat in the factory."}, {"modality": "vision.", "evidence": "At 0:56, the person in a brown suit also wears a tie and has black hair is the same person as the person with the blue hat.", "inference": "The person in a brown suit is the person with a blue hat."}, {"modality": "vision", "evidence": "At 3:58, the person in brown bows and apologizes after bumping into people on the way to find his father.", "inference": "He bumped into 1 person."}, {"modality": "audio", "evidence": "At 3: 58, a sound of bumping can be heard.", "inference": "Someone bumps into other one."}, {"modality": "vision", "evidence": "At 4: 01, the person in brown bumps into a person with grey hair, then bows and apologizes.", "inference": "He bumped into the second person while looking for his father."}, {"modality": "vision", "evidence": "At 4: 04, when the person in brown hair turns back to see his partner, the person in brown suit is bowing to an old man whose top of the head is bald.", "inference": "He bumped into the third person."}, {"modality": "vision", "evidence": "At 4: 08, the person in brown suit is bowing to a person holding a computer bag in hand.", "inference": "He bumped into the fourth person."}, {"modality": "audio", "evidence": "A man's voice called 'Dad.'", "inference": "He find his father."}, {"modality": "vision", "evidence": "At 4:11, a person with white hair appears in the train. It is his father.", "inference": "The person with brown hair find the father."}]}
{"id": "video_511__q0000", "video_id": "video_511", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_511_t5s.mp4"], "extra_info": {"question": "Who is free from danger right now?"}, "reward_model_ground_truth": {"multi_choice": ["A. the cat", "B. Svetlana", "C. Sussman", "D. Ki Chi"], "correct_option": "C"}, "answer_text": "Sussman", "reasoning_steps": [{"modality": "vision", "evidence": "The cat's face has 'sucked into itself' due to a pressure change.", "inference": "The cat is not free from danger right now."}, {"modality": "vision", "evidence": "At 4:31 Ki chi has already 'hit a wall and shattered'.", "inference": "Ki chi is not free from danger right now."}, {"modality": "audio", "evidence": "At 3:43, the man ordered Sussman to look for Svetlana.", "inference": "Svetlana is not free from danger right now."}, {"modality": "vision", "evidence": "Finally, Sussman is operating the spaceship", "inference": "Sussman is free from danger right now."}]}
{"id": "video_512__q0000", "video_id": "video_512", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_512_t5s.mp4"], "extra_info": {"question": "What is the specific time of the Symposium on Molecular Patronium?"}, "reward_model_ground_truth": {"multi_choice": ["A Tuesday at 6 PM", "B Tuesday at 8 PM", "C Friday at 6 PM", "D Friday at 8 PM"], "correct_option": "A"}, "answer_text": "Tuesday at 6 PM", "reasoning_steps": [{"modality": "audio", "evidence": "A girl's voice starts to sing dramatically.", "inference": "Other people's reactions to this song might have to do with the answer."}, {"modality": "vision", "evidence": "At the same moment, two boys were walking up the stairs.", "inference": "We should observe their reactions to the song."}, {"modality": "audio", "evidence": "At 00:18, the other boy said, 'If cats could sing, they hated it, too.'", "inference": "Both of them didn't like the girl's song."}, {"modality": "audio", "evidence": "At 1:43 seconds , the girl mentions that the time is Friday at 8 PM.", "inference": "Mark Friday at 8 PM as the time for the showcase."}, {"modality": "audio", "evidence": "At 1:52, the glasses guy before said that they are going to the Symposium on Molecular Patronium on Friday.", "inference": "The glasses guy said an event that is contrarary to the showcase schedule."}, {"modality": "audio", "evidence": "At 1:54, the other guy corrected him that the specific time of the symposium is Tuesday at 6 PM.", "inference": "Be careful to identify which one is the really correct time."}, {"modality": "audio", "evidence": "At 1:56, the other boy wearing glasses nervously denied the other boy's words.", "inference": "He's denying it awkwardly because he doesn't like the girl's song and therefore lying. So the specific time of the Symposium on Molecular Patronium is Tuesday at 6 PM."}]}
{"id": "video_513__q0000", "video_id": "video_513", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_513_t5s.mp4"], "extra_info": {"question": "How many people in the scene are looking at their phones when Penny announces she got a part on a TV show?"}, "reward_model_ground_truth": {"multi_choice": ["A 3", "B 4", "C 5", "D 6"], "correct_option": "A"}, "answer_text": "3", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:03, Penny announces she got a part on a TV show.", "inference": "Get the TV show  Penny talked about."}, {"modality": "vision", "evidence": "From vision at 0:06, when she finished saying this, there are 3 people who are looking at cell phones.", "inference": "3 people in the scene are looking at their phones when Penny announces she got a part on a TV show."}]}
{"id": "video_515__q0000", "video_id": "video_515", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_515_t5s.mp4"], "extra_info": {"question": "What is the name of the first pet that the vlogger feeds in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A Ella", "B Lucy", "C Stanley", "D Bud"], "correct_option": "A"}, "answer_text": "Ella", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:23 seconds in the video, the video shows that the vlogger first feeds a rat (kind of dog).", "inference": "This dog was fed by this vlogger."}, {"modality": "audio", "evidence": "The vlogger says, 'now ella gets the dinner.' at 6:19 while the scene focuses on that rat (kind of dog).", "inference": "The name of this dog is Ella."}]}
{"id": "video_516__q0000", "video_id": "video_516", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_516_t5s.mp4"], "extra_info": {"question": "Who did not care what Laz said?"}, "reward_model_ground_truth": {"multi_choice": ["A. S3X PANDA99", "B. Doug", "C. Jake", "D. Mark"], "correct_option": "C"}, "answer_text": "Jake", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:50, Doug and Mark was Laz's colleagues.", "inference": "Doug and Mark cared what Laz said."}, {"modality": "audio", "evidence": "At 3:43,  S3X PANDA99 answered Laz's questions.", "inference": "S3X PANDA99 cared what Laz said."}, {"modality": "audio", "evidence": "At 0:33, Jake did not show up in the TV show, because he took PSATs.", "inference": "Jake did not care what Laz said."}]}
{"id": "video_516__q0001", "video_id": "video_516", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_516_t5s.mp4"], "extra_info": {"question": "After one asked 'Do you want to go backstage', how many new person came with him?"}, "reward_model_ground_truth": {"multi_choice": ["A. 5", "B. 6", "C. 7", "D. 8"], "correct_option": "B"}, "answer_text": "6", "reasoning_steps": [{"modality": "audio", "evidence": "At the 4:06, one asked 'Do you want to go backstage'.", "inference": "This indicates that the fan and her friend were invited to the backstage."}, {"modality": "vision", "evidence": "From 4:17, the girl's six friends were shown.", "inference": "There were 6 new persons came with him."}]}
{"id": "video_517__q0000", "video_id": "video_517", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_517_t5s.mp4"], "extra_info": {"question": "When the music completely changes into a new rhythm for the last time, what is the singer doing?"}, "reward_model_ground_truth": {"multi_choice": ["A: Dancing with handkerchiefs in his hand", "B: Dancing with handkerchiefs in his hand while singing", "C: Only singing without dancing", "D: Standing still and clapping"], "correct_option": "A"}, "answer_text": "Dancing with handkerchiefs in his hand", "reasoning_steps": [{"modality": "vision", "evidence": "The main singer is clearly visible.", "inference": "Identify the main singer is the man with a moustache."}, {"modality": "audio", "evidence": "A clear rhythm change happens late in the performance.", "inference": "Mark the point that need attention to check the singer's actions."}, {"modality": "audio", "evidence": "The singer's voice is not heard at that point.", "inference": "Confirm the singer is not singing during the final rhythm change."}, {"modality": "vision", "evidence": "The singer is dancing and holding handkerchiefs.", "inference": "Show the singer's dancing with handkerchiefs in his hand at that monment."}]}
{"id": "video_518__q0000", "video_id": "video_518", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_518_t5s.mp4"], "extra_info": {"question": "Who was the last person hit by the huge guy?"}, "reward_model_ground_truth": {"multi_choice": ["A Jeremy", "B Jojo", "C Nick", "D Alec"], "correct_option": "C"}, "answer_text": "Nick", "reasoning_steps": [{"modality": "vision", "evidence": "At 4:23, the huge guy used his fingers to bounce a pink hair boy away.", "inference": "He might be the  last person hit by the huge guy."}, {"modality": "audio", "evidence": "At 4:24, the narrator said Nick goes to 'sleep'.", "inference": "Nick was lying down because he's hit by the huge guy."}, {"modality": "vision", "evidence": "At the same moment, the pink hair boy was sleeping with a drop of blood flowing out of his mouth.", "inference": "Initially confrim that pink hair boy is Nick."}, {"modality": "vision", "evidence": "At 4:43 seconds in the video，Alec and the huge guy fought.", "inference": "If the huge guy was beaten, then Nick would be the last person that he hit."}, {"modality": "vision", "evidence": "At 4:50, the huge guy was beaten.", "inference": "Confirm that Nick was the last person that was hit by him."}]}
{"id": "video_519__q0000", "video_id": "video_519", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_519_t5s.mp4"], "extra_info": {"question": "How many knives are there on the knife rack when Peggy's husband talks to Peggy? "}, "reward_model_ground_truth": {"multi_choice": ["A 4", "B 5", "C 6", "D 7"], "correct_option": "B"}, "answer_text": "5", "reasoning_steps": [{"modality": "audio", "evidence": "Peggy says to her son, 'If your father finds out,' indicating that the man is the little boy's father, Peggy's husband.", "inference": "The man who gets into the kitchen is Peggy's husband."}, {"modality": "vision", "evidence": "when Peggy's husband finds the little girl and talks to Peggy in the kitchen, there are 5 knives on the knife rack.", "inference": "Five knives are on the knife rack when Peggy's husband talks to Peggy."}]}
{"id": "video_52__q0000", "video_id": "video_52", "qid": 0, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_52_t5s.mp4"], "extra_info": {"question": "What is making noise inside the car while driving?"}, "reward_model_ground_truth": {"multi_choice": ["A. A dolly.", "B. A parcel.", "C. The car door.", "D. Carriage connection."], "correct_option": "B"}, "answer_text": "A parcel.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:01, the driver is driving the car.", "inference": "Pay attention to the noise inside the car during the driving process."}, {"modality": "audio", "evidence": "At 00:04, there is noise inside the car.", "inference": "Pay attention to things on the car that could make such noise."}, {"modality": "vision", "evidence": "At 00:18, the driver stop the bus and take down a parcel.", "inference": "The noise is from the parcel."}]}
{"id": "video_52__q0001", "video_id": "video_52", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_52_t5s.mp4"], "extra_info": {"question": "Where is the second destination in relation to the first?"}, "reward_model_ground_truth": {"multi_choice": ["A. On the street behind.", "B. On the street to the left.", "C. On the street to the right.", "D. Across the street."], "correct_option": "D"}, "answer_text": "Across the street.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:25, the driver arrives at the first destination.", "inference": "Pay attention to the next destination's place."}, {"modality": "audio", "evidence": "At 00:46, the driver says he makes a full U turn.", "inference": "The car need to make a U turn to arrive."}, {"modality": "vision", "evidence": "At 00:55, the car goes straight and stop after making the U turn.", "inference": "Across the street from the first destination,  there is the second."}]}
{"id": "video_520__q0000", "video_id": "video_520", "qid": 0, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_520_t5s.mp4"], "extra_info": {"question": "When the boy said 'Ok' for the second time, what time was it?"}, "reward_model_ground_truth": {"multi_choice": ["A 11:02", "B 12:55", "C 11:33", "D 6:59"], "correct_option": "C"}, "answer_text": "11:33", "reasoning_steps": [{"modality": "audio", "evidence": "At the 0:59 timestamp, the boy says 'Ok' for the first time.", "inference": "The boy said 'Ok' for the first time."}, {"modality": "vision", "evidence": "At the 1:07 timestamp, the clock changed from 11:02 to 11:33.", "inference": "It's 11:33."}, {"modality": "audio", "evidence": "At the 1:22 timestamp, the boy says 'Ok' for the second time.", "inference": "The boy said 'Ok' for the second time."}]}
{"id": "video_521__q0000", "video_id": "video_521", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_521_t5s.mp4"], "extra_info": {"question": "In what way does the music that is being played with the vinyl player match the scene's visuals and content?"}, "reward_model_ground_truth": {"multi_choice": ["A. It adds a comical or sarcastic tone.", "B. It matches the sadness of the main character.", "C. It adds to the calm and moody ambience of the scene.", "D. It plays thunderstorms sounds, matching the environmental sounds."], "correct_option": "A"}, "answer_text": "It adds a comical or sarcastic tone to what is otherwise a sad scene.", "reasoning_steps": [{"modality": "vision", "evidence": "At the 2:21, the scene is dark, with the characters having primarily sad facial expressions.", "inference": "the content is likely sad."}, {"modality": "vision", "evidence": "At the 2:28, the male lead turns around but bumps into the glass door, which reveals his inner sorrow and distraction.", "inference": "The scene exuding a somewhat comical melancholy."}, {"modality": "audio", "evidence": "At the 2:21, the vinyl plays a melodious and sorrowful song.", "inference": "The music is  consistent with the mood of the scene, thus adding a comical or sarcastic tone to it."}]}
{"id": "video_522__q0000", "video_id": "video_522", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_522_t5s.mp4"], "extra_info": {"question": "Why are the audience laughing when the four people looked up from the table?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because the coin drops onto the table.", "B.Because his movement is very exaggerated.", "C.Because others ignore him.", "D.Because he ignores the rules."], "correct_option": "D"}, "answer_text": "Because he ignores the rules.", "reasoning_steps": [{"modality": "vision", "evidence": "from vision at 3:06, the four people looked up from the table.", "inference": "Identify the time the four people looked up."}, {"modality": "audio", "evidence": "from audio at 3:06, Ross says that the coin determines who starts first.", "inference": "Ross forgets to say 'somebody calls this time'."}, {"modality": "audio", "evidence": "The audience are laughing.", "inference": "Because he ignores the rules."}]}
{"id": "video_524__q0000", "video_id": "video_524", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_524_t5s.mp4"], "extra_info": {"question": "What is the state of the chessboard when the man in the white says 'Checkmate'?"}, "reward_model_ground_truth": {"multi_choice": ["A. The white rook and queen are attacking the black king.", "B. The white knight and queen are attacking the black king.", "C. The white pawn and bishop deliver the checkmate.", "D. The white king delivers the checkmate."], "correct_option": "D"}, "answer_text": "The white king delivers the checkmate.", "reasoning_steps": [{"modality": "audio", "evidence": "At the 4:51, the man in the white shirt says 'Checkmate'.", "inference": "The scene of the chessboard is displayed just before and after this very moment."}, {"modality": "vision", "evidence": "At the 4:50, the chessboard shows that the black king attacks.", "inference": "The white king is in checkmate."}]}
{"id": "video_525__q0000", "video_id": "video_525", "qid": 0, "question_type": "temporal understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_525_t5s.mp4"], "extra_info": {"question": "At what point in the performance do the audience stand up?"}, "reward_model_ground_truth": {"multi_choice": ["A: When the performance ends", "B: When the performance starts", "C: Before the performance ends", "D: The audience do not stand up"], "correct_option": "C"}, "answer_text": "Before the performance ends", "reasoning_steps": [{"modality": "vision", "evidence": "At 4:03, audience members are standing during the performance.", "inference": "Identifies when the standing occurs."}, {"modality": "audio", "evidence": "Singers are still performing and the performance has not ended yet.", "inference": "Standing happens before the performance ends."}]}
{"id": "video_526__q0000", "video_id": "video_526", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_526_t5s.mp4"], "extra_info": {"question": "What are the things for raising a dog mentioned in order?"}, "reward_model_ground_truth": {"multi_choice": ["A. Food, fence, nylon rope, dog bowl (stainless steel), toys", "B. Food, fence, dog bowl (iron steel), nylon rope, toys", "C. Food, fence,  dog bowl (stainless steel), nylon rope, toys", "D. Food, fence, nylon rope, dog bowl (iron steel), toys"], "correct_option": "A"}, "answer_text": "Food, fence, nylon rope, dog bowl (stainless steel), toys", "reasoning_steps": [{"modality": "vision", "evidence": "The video shows a dog food in time 0:46.", "inference": "The first things for raising a dog is food."}, {"modality": "vision", "evidence": "The video shows a fence in time 1:46.", "inference": "The second things for raising a dog is fence."}, {"modality": "vision", "evidence": "The video shows a nylon rope in time 1:58.", "inference": "The third things for raising a dog is nylon rope."}, {"modality": "audio", "evidence": "The man says stainless steel dog bowl in time 3:05.", "inference": "The fourth things for raising a dog is stainless steel dog bowl."}, {"modality": "audio", "evidence": "The man says toys in time 3:35.", "inference": "The fifth things for raising a dog is toys."}]}
{"id": "video_527__q0000", "video_id": "video_527", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_527_t5s.mp4"], "extra_info": {"question": "If the user wanted the Digispark to automatically accept the UAC prompt for the Command Prompt, which modification to the 'test.ino' sketch would be necessary?"}, "reward_model_ground_truth": {"multi_choice": ["A. Change `DigiKeyboard.sendKeyStroke(KEY_ARROW_LEFT);` to `DigiKeyboard.sendKeyStroke(KEY_ARROW_RIGHT);`.", "B. Remove the line `DigiKeyboard.sendKeyStroke(KEY_ARROW_LEFT);` entirely.", "C. Add `DigiKeyboard.sendKeyStroke(KEY_Y);` after `DigiKeyboard.print('cmd');`.", "D. Change `DigiKeyboard.sendKeyStroke(KEY_ENTER);` to `DigiKeyboard.sendKeyStroke(KEY_SPACE);`."], "correct_option": "B"}, "answer_text": "Remove the line `DigiKeyboard.sendKeyStroke(KEY_ARROW_LEFT);` entirely.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:22, the man said' remnant mode prompts and we will get the answer.'", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "A few rows above it, there are some codes that need to be run.", "inference": "after remove the line `DigiKeyboard.sendKeyStroke(KEY_ARROW_LEFT);` entirely, it will run  automatically."}]}
{"id": "video_528__q0000", "video_id": "video_528", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_528_t5s.mp4"], "extra_info": {"question": "What instrument playing technique does the lead dancer imitate at the start of the song with the cloth movements?"}, "reward_model_ground_truth": {"multi_choice": ["A. Vibrato", "B. Tremolo", "C. Staccato", "D. Trill"], "correct_option": "D"}, "answer_text": "Trill", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:35, the man at the start of the line when the dance group comes in takes distinct actions.", "inference": "They are likely the leader."}, {"modality": "vision", "evidence": "At 0:52, the leader shakes the two cloths he is holding in place.", "inference": "This is likely the imitation action."}, {"modality": "audio", "evidence": "At that time, the instrument plays notes with pitch oscillation, which can be considered as matching the shaking movement of the cloths.", "inference": "Confirms it's the instrument playing technique in question."}, {"modality": "audio", "evidence": "The pitch oscillation is at least half a semitone.", "inference": "Confirms it's a trill rather than vibrato."}]}
{"id": "video_529__q0000", "video_id": "video_529", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_529_t5s.mp4"], "extra_info": {"question": "What kind of attitude this person who was considered a Twitter commentator by the anchor commented on the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. Moderate.", "B. Criticized.", "C. Praised.", "D. Angry."], "correct_option": "C"}, "answer_text": "Praised.", "reasoning_steps": [{"modality": "audio", "evidence": "From audio, he talked about how he felt that someone was a Twitter commentator at around 4:55.", "inference": "Get theTwitter commentator he talked about."}, {"modality": "vision", "evidence": "From vision, the comment made by this Twitter commentator reads:Kanye's voice just feels so real in this.", "inference": "This is a praised attitude to the video."}]}
{"id": "video_53__q0000", "video_id": "video_53", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_53_t5s.mp4"], "extra_info": {"question": "What does the mother call her daughter when she asks about her makeup?"}, "reward_model_ground_truth": {"multi_choice": ["A. Baby.", "B. Yu Xi.", "C. Daughter.", "D. Xixi."], "correct_option": "D"}, "answer_text": "Xixi.", "reasoning_steps": [{"modality": "vision", "evidence": "At 5:10,the mother walks into her daughter's room when she is doing makeup.", "inference": "locate this point."}, {"modality": "audio", "evidence": "the mom said 'xixi,look,have I improved?'.", "inference": "Xixi."}]}
{"id": "video_530__q0000", "video_id": "video_530", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_530_t5s.mp4"], "extra_info": {"question": "After the coach demonstrates the dribble move, what is the specific requirement he explains for the very next training exercise?"}, "reward_model_ground_truth": {"multi_choice": ["A. A slam dribble, then between-the-legs, then behind-the-back.", "B. A crossover wider than your knees while shifting weight.", "C. Omit the slam dribble, only do a back-and-forth joke cross.", "D. A split, then a slam dribble and between-the-legs."], "correct_option": "C"}, "answer_text": "Omit the slam dribble, only do a back-and-forth joke cross.", "reasoning_steps": [{"modality": "vision", "evidence": "At timestamp 01:27, the text overlay 'JOKE CROSS' appears on the screen.", "inference": "The user must first identify the segment."}, {"modality": "vision", "evidence": "The coach  performs a 'pound to joke cross' drill, which involves one high pound dribble followed by a quick, low crossover (the joke cross).", "inference": "The user needs to watch the drill being demonstrated at this time."}, {"modality": "audio", "evidence": "Starting at 02:06, the coach says, 'Now I just want no pounds. So, take the pound away, and I want a joke cross back and forth.'", "inference": "The user must then listen to the coach's instructions immediately following the completion of the 'pound to joke cross' drill."}, {"modality": "vision", "evidence": "The coach then demonstrates this new drill, which consists of continuous low joke crosses without any preceding pound dribble.", "inference": "The answer is 'Omit the slam dribble and only do a back-and-forth joke cross dribble.'"}]}
{"id": "video_531__q0000", "video_id": "video_531", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_531_t5s.mp4"], "extra_info": {"question": "Why are the audience laughing when the someone says it's not his mom"}, "reward_model_ground_truth": {"multi_choice": ["A.Because Seldon confuses Bernadette with his mom.", "B.Because Bernadette's sounds similar to his mom.", "C.Because Seldon thinks it's very unsettling.", "D.Because Bernadette askes the audiences's suggestion."], "correct_option": "B"}, "answer_text": "Because Bernadette's sounds similar to his mom.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:07, a woman asked who is it, and her voice is hoarse.", "inference": "The sounds similar to Howard's mom."}, {"modality": "audio", "evidence": "At 0:13, Howard says 'It's not my mother, it's Bernadette'.", "inference": "The audience maybe are laughing for Bernadette likes Howard's mother afer they made couple."}, {"modality": "vision", "evidence": "At 0:20, Bernadatte came out.", "inference": "The lady who shouted before is indeed Bernadatte."}, {"modality": "audio", "evidence": "Bernadette's voice is sweet.", "inference": "A contrast is created, her hoarse sounds similar to his mom."}]}
{"id": "video_532__q0000", "video_id": "video_532", "qid": 0, "question_type": "temporal understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_532_t5s.mp4"], "extra_info": {"question": "What action does the cook take when the second background music track starts?"}, "reward_model_ground_truth": {"multi_choice": ["A. He removes the insides of the pumpkin.", "B. He adds the rice to the skillet.", "C. He adds the insides of the pumpkin to the skillet.", "D. He stirs vegetables in the skillet."], "correct_option": "B"}, "answer_text": "He adds the rice to the skillet.", "reasoning_steps": [{"modality": "audio", "evidence": "The background music  track changes at timestamp 3:18.", "inference": "This is the point where the action needs to be identified."}, {"modality": "vision", "evidence": "At that point, the cook adds rice into the skillet.", "inference": "The cook adds the rice to the skillet."}]}
{"id": "video_533__q0000", "video_id": "video_533", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_533_t5s.mp4"], "extra_info": {"question": "What color is Drew's own clothing?"}, "reward_model_ground_truth": {"multi_choice": ["A Red", "B Blue", "C Green", "D White"], "correct_option": "B"}, "answer_text": "Blue", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:37, the protagonist praises a Starbucks employee named Drew.", "inference": "After this point, the people: Drew mentioned in the question will be shown."}, {"modality": "vision", "evidence": "At 3:46, Drew is shown, wearing a blue short-sleeved shirt under a green Starbucks apron.", "inference": "The color of Drew's own clothing is blue."}]}
{"id": "video_534__q0000", "video_id": "video_534", "qid": 0, "question_type": "summarization", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_534_t5s.mp4"], "extra_info": {"question": "When a traditional song of the region starts being played, what scene is initially shown?"}, "reward_model_ground_truth": {"multi_choice": ["A. A drone shot of the village.", "B. Two women in traditional clothing being interviewed.", "C. A woodworking workshop.", "D. A drone shot of a man."], "correct_option": "A"}, "answer_text": "A drone shot of the village.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:03, the video is situated in a village, possibly at a mountainous region in Greece.", "inference": "Likely a Greek village."}, {"modality": "audio", "evidence": "At 0:40, the Greek language is spoken.", "inference": "Confirms a Greek village."}, {"modality": "audio", "evidence": "At 8:27, there are mentions and usage of the Vlach language.", "inference": "Confirms it's a Greek village with Vlach cultural roots."}, {"modality": "audio", "evidence": "At timestamp 10:11, a distinct song with different instruments starts.", "inference": "Likely the song in question."}, {"modality": "audio", "evidence": "The song contains the clarinet, a drum, and a rattle percussion, which is traditional of Vlach and some Greek music", "inference": "Confirms this is the song in question."}, {"modality": "vision", "evidence": "At the time 10:12, the song starts playing and a drone shot of the village is shown.", "inference": "This is the answer."}]}
{"id": "video_535__q0000", "video_id": "video_535", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_535_t5s.mp4"], "extra_info": {"question": "When the man mentions 'Assemble a sales team', how many humanoid trophies are on the second shelf of the rack behind him?"}, "reward_model_ground_truth": {"multi_choice": ["A 3", "B 4", "C 5", "D 6"], "correct_option": "B"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:16, the man mentions 'Assemble a sales team'.", "inference": "Get the trophy rack they talked about."}, {"modality": "vision", "evidence": "From vision, the second shelf of the trophy rack has 4 humanoid trophies.", "inference": "There are 4 humanoid trophieson on the second shelf of the rack behind him."}]}
{"id": "video_538__q0000", "video_id": "video_538", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_538_t5s.mp4"], "extra_info": {"question": "What time could it be when the blogger introduces themselves in the car?"}, "reward_model_ground_truth": {"multi_choice": ["A 1:39", "B 2:39", "C 3:39", "D 4:39"], "correct_option": "A"}, "answer_text": "1:39", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:22, in the car, the blogger introduced that she's going to do a school shopping to do a give-away activity in Target.", "inference": "Set the point to check the vision information."}, {"modality": "vision", "evidence": "At the same moment, the sky and environment light outside the car is bright.", "inference": "So when the blogger was in the car, the time is stil in daytime."}, {"modality": "audio", "evidence": "At 05:54, the blogge said, 'All right, guys, now it's 2:39. Which means I have been here for an hour.'", "inference": "Make the time backwards for an hour is 1:39, which is the closest time when the blogger was in the car."}]}
{"id": "video_539__q0000", "video_id": "video_539", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_539_t5s.mp4"], "extra_info": {"question": "What time signature is the second music piece that is played as part of the shadow theater performance?"}, "reward_model_ground_truth": {"multi_choice": ["A. 7/8", "B. 4/4", "C. 3/4", "D. 8/8"], "correct_option": "A"}, "answer_text": "7/8", "reasoning_steps": [{"modality": "vision", "evidence": "The performer goes behind the screen at 0:46.", "inference": "This is likely when the play starts."}, {"modality": "vision", "evidence": "The performer comes out from behind the screen at 6:08 .", "inference": "Confirm the duration of the second music piece."}, {"modality": "audio", "evidence": "The second song has a rhythmic pattern of 3-2-2.", "inference": "The song's time signature is 7/8."}]}
{"id": "video_54__q0000", "video_id": "video_54", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_54_t5s.mp4"], "extra_info": {"question": "What is the person with red hair doing when the person with green hair says something wrong?"}, "reward_model_ground_truth": {"multi_choice": ["A. Sit on the sofa,looking at the person with green hair.", "B. Hug one's legs,looking at the person with green hair.", "C. Squat down to watch the bugs.", "D. With hands support the head and lean against it."], "correct_option": "D"}, "answer_text": "With hands support the head and lean against it.", "reasoning_steps": [{"modality": "vision", "evidence": "At the timestamp 0:38, there are many people in the living room.", "inference": "The person with red hair and the person with green hair are both in the scene."}, {"modality": "audio", "evidence": "At the timestamp 0:43,the person with green hair says,'there is no food'.", "inference": "The person with green hair says something wrong."}, {"modality": "vision", "evidence": "The person with red hair is leaning against the sofa.", "inference": "The person with red hair is leaning against the sofa when the person with green hair says something wrong."}, {"modality": "vision", "evidence": "The person  with hands support the head.", "inference": "The person with red hair is leaning against the sofa with hands support the head when the person with green hair says something wrong."}]}
{"id": "video_540__q0000", "video_id": "video_540", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_540_t5s.mp4"], "extra_info": {"question": "What are the color of the avatars in the group that says 'you guys want to play again' during the second entry?"}, "reward_model_ground_truth": {"multi_choice": ["A Purple", "B Blue", "C Red", "D Green"], "correct_option": "A"}, "answer_text": "Purple", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:27, the one who says 'you guys want to play again' is a woman.", "inference": "Now we confirm the target is the woman."}, {"modality": "vision", "evidence": "At 0:32, the team alejandor is shown and there's only one female avatar which is the first purple one.", "inference": "The color of the avatars in the question is purple."}]}
{"id": "video_543__q0000", "video_id": "video_543", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_543_t5s.mp4"], "extra_info": {"question": "What is the third thing that the vlogger does after she said 'One of my favorite features on this app...'?"}, "reward_model_ground_truth": {"multi_choice": ["A: The vlogger removes extra hair.", "B: The vlogger removes a walking pedestrian.", "C: The vlogger removes Number 17.", "D: The vlogger removes a Hello Kitty sculpture."], "correct_option": "C"}, "answer_text": "The vlogger removes Number 17.", "reasoning_steps": [{"modality": "audio", "evidence": "The vlogger said that at 6:37.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "The vlogger first removes extra hair at 6:42 of the video.", "inference": "Removing the extra hair is the first thing."}, {"modality": "vision", "evidence": "The vlogger then removes a walking pedestrian at 6:47 of the video.", "inference": "Removing the walking pedestrian is the second thing."}, {"modality": "vision", "evidence": "The vlogger then removes Number 17 at 6:50 of the video.", "inference": "Removing Number 17 is the third thing."}]}
{"id": "video_545__q0000", "video_id": "video_545", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_545_t5s.mp4"], "extra_info": {"question": "When the main character delivers the line '[...] you are the real murderer, no doubt' in English, what scene immediately follows?"}, "reward_model_ground_truth": {"multi_choice": ["A. The camera zooms into the accusee's reaction", "B. The camera zooms into another man's reaction to the statement", "C. The camera pans to the reaction of several men in the table", "D. The camera zooms into the accuser's reaction"], "correct_option": "A"}, "answer_text": "The camera zooms into the accusee's reaction", "reasoning_steps": [{"modality": "audio", "evidence": "The primarily language of the show is Tami.", "inference": "There might be a few moments the characters use English."}, {"modality": "vision", "evidence": "After the accusation, the camera changes from pointing on the accuser to zooming into the accusee's reaction.", "inference": "Confirms the scene that succeeds the accusation contains the accusee's reaction."}]}
{"id": "video_546__q0000", "video_id": "video_546", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_546_t5s.mp4"], "extra_info": {"question": "What is the first music note that the instructor plays, excluding the prerecorded music and samples he uses?"}, "reward_model_ground_truth": {"multi_choice": ["A. E3", "B. E2", "C. C3", "D. C2"], "correct_option": "C"}, "answer_text": "C3", "reasoning_steps": [{"modality": "vision", "evidence": "The piano roll tab is open.", "inference": "Confirms this is a relevant instance."}, {"modality": "vision", "evidence": "The cursor hovers over the C3 note, and the note changes color.", "inference": "This is likely the correct note."}, {"modality": "audio", "evidence": "The first musical note that can be heard is C3 played multiple times.", "inference": "C3."}]}
{"id": "video_548__q0000", "video_id": "video_548", "qid": 0, "question_type": "temporal understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_548_t5s.mp4"], "extra_info": {"question": "When the singer first begins singing, a Tar is already playing. After a Kamancheh player joins, the first Tar stops while the Kamancheh continues. Which musician joins next?"}, "reward_model_ground_truth": {"multi_choice": ["A: The same Tar player who was playing at the start", "B: A different Tar player", "C: A Setar player", "D: A different Kamancheh player"], "correct_option": "B"}, "answer_text": "A different Tar player", "reasoning_steps": [{"modality": "audio", "evidence": "At the start of the singer's vocals, a Tar is already being played by one musician.", "inference": "Establishes the first instrument with the singer."}, {"modality": "audio", "evidence": "A Kamancheh player begins playing after the singer starts.", "inference": "Adds the second instrument."}, {"modality": "audio", "evidence": "The first Tar sound fades out while the Kamancheh continues.", "inference": "Marks when the initial Tar stops."}, {"modality": "vision", "evidence": "Another Tar player, different from the first, is shown playing.", "inference": "A different Tar player."}]}
{"id": "video_55__q0000", "video_id": "video_55", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_55_t5s.mp4"], "extra_info": {"question": "What is the fourth animal seen in the Safari Zone?"}, "reward_model_ground_truth": {"multi_choice": ["A.Brown Bear", "B.Siberian Tiger", "C.Jackal", "D.cheetah"], "correct_option": "C"}, "answer_text": "Jackal", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:50, a woman said 'We are going to the Safari Zone'.", "inference": "This indicates that they would go to see the animals in the Safari Zone afterwards."}, {"modality": "vision", "evidence": "At 5:08, a brown bear appear in the scene.", "inference": "This is the first animal seen in the Safari Zone."}, {"modality": "vision", "evidence": "At 5:32, several siberian tigers appears in the scene.", "inference": "This is the second animal seen in the Safari Zone."}, {"modality": "vision", "evidence": "At 5:53, two white tigers appear in the scene.", "inference": "This is the third animal seen in the Safari Zone."}, {"modality": "vision", "evidence": "At 6:04, a jackal appear in the scene.", "inference": "This is the fourth animal seen in the Safari Zone. So the answer is 'Jackal'."}]}
{"id": "video_550__q0000", "video_id": "video_550", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_550_t5s.mp4"], "extra_info": {"question": "Why dows this person dress up as a pie?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because Gnomes love pie more than anything else in the world.", "B. Because he wanted to take the Gnomes' blackberries.", "C. Because he planned to tease and annoy the Gnomes.", "D. Because he hoped to attract the attention of the Gnomes."], "correct_option": "B"}, "answer_text": "Because he wanted to take the Gnomes' blackberries.", "reasoning_steps": [{"modality": "vision", "evidence": "The person dress up as a pie comes out at the beginning of the video.", "inference": "Identify the right person."}, {"modality": "audio", "evidence": "A yellow hair girl asks 'Why you dress up as a pie ?' .", "inference": "She asks the reason why the person dress up as a pie."}, {"modality": "audio", "evidence": "The person dress up as a pie says the Gnomes loves the pie most.", "inference": "He dressed up as a pie because Gnomes love it ."}, {"modality": "audio", "evidence": "The person says every time Mr.Gnome is about to eat the blackberry he will run in front of him dressed up as a pie.", "inference": "He wants to take the Gnomes' blackberries."}]}
{"id": "video_550__q0001", "video_id": "video_550", "qid": 1, "question_type": "spatial understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_550_t5s.mp4"], "extra_info": {"question": "Where do something they have consulted fall?"}, "reward_model_ground_truth": {"multi_choice": ["A. Fell onto a red car right at the start.", "B. Landed in the hands of two winged elves.", "C. Dropped straight into the Gnomes' open mouth.", "D. Fell into the hands of a girl with purple hair."], "correct_option": "A"}, "answer_text": "Fell onto a red car right at the start.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:16,they consult blackberry", "inference": "The thing is blackberry"}, {"modality": "vision", "evidence": "At 0:43, the Gnome holds a blackberry in hand.", "inference": "Locate the blackberry."}, {"modality": "vision", "evidence": "The person dress up as a pie comes out at 0:47.", "inference": "This person tried to attract the Gnome."}, {"modality": "vision", "evidence": "Then Gnome drops the  blackberry at 0:50.", "inference": "The blackberry will fall down."}, {"modality": "vision", "evidence": "Then two winged elves standing on a red car catch the blackberry.", "inference": "The blackberry fell onto a red car at the start."}]}
{"id": "video_550__q0002", "video_id": "video_550", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_550_t5s.mp4"], "extra_info": {"question": "What is the correct sequence of events that results in the elves collecting the blackberries?"}, "reward_model_ground_truth": {"multi_choice": ["A. Mr. Gnome attempts to eat a blackberry but gets distracted by something else, then he finds another blackberry, which leads to the 'pie' running in front of him to distract him, allowing the elves to collect the berries.", "B. Mr. Gnome first spots a blackberry, then the 'pie' immediately runs to distract him before he can eat it, and then the elves gather the berries.", "C. The character in the 'pie' costume explains his plan, then the elves collect the blackberries, and finally, Mr. Gnome tries to eat a blackberry.", "D. Mr. Gnome is first distracted by the 'pie' costume, then tries to eat a blackberry unsuccessfully, and the elves collect the berries."], "correct_option": "A"}, "answer_text": "Mr. Gnome attempts to eat a blackberry but gets distracted by something else, then he finds another blackberry, which leads to the 'pie' running in front of him to distract him, allowing the elves to collect the berries.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:05, professor Wise said ' Mr. Gnome is eating blackberries.'", "inference": "get the point."}, {"modality": "vision", "evidence": "The man who is eating blackberries is Mr. Gnome.", "inference": "get the Mr. Gnome."}, {"modality": "vision", "evidence": "when Mr. Gnome wants to eat another blackberry,he is distracted by a 'pie'.", "inference": "get the thing firstly."}, {"modality": "vision", "evidence": "the children are collecting the berries.", "inference": "it happened sequently."}]}
{"id": "video_550__q0003", "video_id": "video_550", "qid": 3, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_550_t5s.mp4"], "extra_info": {"question": "What is the stated and demonstrated main purpose of the character dressed as a pie?"}, "reward_model_ground_truth": {"multi_choice": ["A. To entertain Mr. Gnome with a costume.", "B. To get Mr. Gnome to share his blackberries.", "C. To prevent Mr. Gnome from eating blackberries.", "D. To help Mr. Gnome find more ripe blackberries."], "correct_option": "C"}, "answer_text": "To prevent Mr. Gnome from eating blackberries.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:05, the man said ' I am dressed as a pie.'", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "the man is dressed as a pie.", "inference": "the man is professor Wise."}, {"modality": "audio", "evidence": "At 0:05, professor Wise said ' Mr. Gnome is eating blackberries.'", "inference": "get the point."}, {"modality": "vision", "evidence": "professor Wise stands in front of Mr. Gnome.", "inference": "maybe he wants to attract Mr. Gnome's attentions."}, {"modality": "vision", "evidence": "after that, the children is collecting the blackberries.", "inference": "To prevent Mr. Gnome from eating blackberries."}]}
{"id": "video_551__q0000", "video_id": "video_551", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_551_t5s.mp4"], "extra_info": {"question": "How many times did Blue superdog use the wrench?"}, "reward_model_ground_truth": {"multi_choice": ["A.7", "B.8", "C.9", "D.10"], "correct_option": "C"}, "answer_text": "9", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:21, Blue superdog brings a wrench.", "inference": "Mark the beginning of the counting."}, {"modality": "audio", "evidence": "At 0.22, the sound effect with wrench appeared three times.", "inference": "There're three times."}, {"modality": "audio", "evidence": "At 0:25, the sound effect with wrench appeared six times.", "inference": "There're nine times in total."}]}
{"id": "video_551__q0001", "video_id": "video_551", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_551_t5s.mp4"], "extra_info": {"question": "What is the collaborative action between Blue Superdog and Dodo to fix the car's wheel?"}, "reward_model_ground_truth": {"multi_choice": ["A. Blue Superdog hands the wheel to Dodo, who then independently attaches it with a tool.", "B. Blue Superdog holds the wheel while Dodo uses a wrench to tighten the screw.", "C. Blue Superdog uses a wrench to tighten the screw while Dodo holds the wheel steady.", "D. Blue Superdog and Dodo both hold the wrench and tighten the screw together."], "correct_option": "C"}, "answer_text": "Blue Superdog uses a wrench to tighten the screw while Dodo holds the wheel steady.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:06, someone said 'Dodo'.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "there is a yellow dog.", "inference": "the yellow dog is 'Dodo'."}, {"modality": "vision", "evidence": "at 0:21, the Blue Superdog uses a wrench to tighten the screw.", "inference": "Blue Superdog uses a wrench to tighten the screw."}, {"modality": "vision", "evidence": "at 0:26, 'Dodo' holds the wheel steady.", "inference": "'Dodo' holds the wheel steady."}]}
{"id": "video_551__q0002", "video_id": "video_551", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_551_t5s.mp4"], "extra_info": {"question": "What action did Dodo take when he said 'OK' for the first time?"}, "reward_model_ground_truth": {"multi_choice": ["A.Squatting with the right arm raised", "B.Sitting with the left arm raised", "C.Squatting with the left arm raised", "D.Sitting with the right arm raised"], "correct_option": "C"}, "answer_text": "Squatting with the left arm raised", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:15, Dodo says 'OK' for the first time.", "inference": "Mark the very point of the question that needs attention."}, {"modality": "vision", "evidence": "At 0:15,  Dodo's squatting with the left arm raised.", "inference": "This is the action at that moment mentioned in the question."}]}
{"id": "video_552__q0000", "video_id": "video_552", "qid": 0, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_552_t5s.mp4"], "extra_info": {"question": "One hasn't laughed,who?"}, "reward_model_ground_truth": {"multi_choice": ["A. The boy with orange-colored hair.", "B. The boy with black-colored hair.", "C. The girl with orange-colored hair.", "D. The girl with blonde-colored hair."], "correct_option": "D"}, "answer_text": "The girl with blonde-colored hair.", "reasoning_steps": [{"modality": "audio", "nevidence": "Laughter can be heard from the point of 20 seconds.", "inference": "Existential laughter"}, {"modality": "vision", "evidence": "At 28 seconds, it can be observed that the blonde girl is not laughing and is a little angry.", "inference": "The blonde girl didn't laugh."}]}
{"id": "video_553__q0000", "video_id": "video_553", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_553_t5s.mp4"], "extra_info": {"question": "What did the girl drawn on the whiteboard say?"}, "reward_model_ground_truth": {"multi_choice": ["A. Persuade another girl not to move.", "B. Narrate how to lock Aether down.", "C. Introduce a good fried radish stand nearby.", "D. Talk about buying new equipment with Yelan."], "correct_option": "A"}, "answer_text": "Persuade another girl not to move.", "reasoning_steps": [{"modality": "vision", "evidence": "A girl is drawing on the whiteboard at 0:03.", "inference": "Locate the girl."}, {"modality": "vision", "evidence": "The girl's brush points at the face of the girl with green hair at 0:04.", "inference": "The girl wants to draw on the face of the girl with green hair."}, {"modality": "audio", "evidence": "Then girl says 'Please stop moving' to the girl with green hair.", "inference": "The girl persuades another girl not to move."}]}
{"id": "video_553__q0001", "video_id": "video_553", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_553_t5s.mp4"], "extra_info": {"question": "Where is the 'NO ONE FIGHT ALONE' poster relative to the person who deployed Culein-Anbar?"}, "reward_model_ground_truth": {"multi_choice": ["A. On the left wall.", "B. On the back wall", "C. On the right wall.", "D. On the front wall."], "correct_option": "B"}, "answer_text": "On the back wall", "reasoning_steps": [{"modality": "vision", "evidence": "The 'NO ONE FIGHT ALONE' poster shows on the wall at 0:23.", "inference": "Locate the position of the poster."}, {"modality": "audio", "evidence": "The girl with long black hair says ' You locked Aether down good at 0:28.", "inference": "She talks to the girl who locked Aether down."}, {"modality": "audio", "evidence": "The girl with green hair answers ' I'm, fine'.", "inference": "She is the girl who locked Aether down."}, {"modality": "vision", "evidence": "The 'NO ONE FIGHT ALONE' poster shows on the wall behind the girl with green hair at 0:36.", "inference": "The poster on the back wall of the person who locked Aether down."}]}
{"id": "video_554__q0000", "video_id": "video_554", "qid": 0, "question_type": "background&music understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_554_t5s.mp4"], "extra_info": {"question": "Which statement about the brief gasp sound is correct?"}, "reward_model_ground_truth": {"multi_choice": ["A. The baby's crying causes the mother to gasp.", "B. The baby's crying causes the parents to gasp.", "C. The baby being pretended as someone else's child causes the mother to gasp.", "D. The baby being pretended as someone else's child causes the parents to gasp."], "correct_option": "B"}, "answer_text": "The baby's crying causes the parents to gasp.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0: 07, the baby is crying.", "inference": "the baby is crying."}, {"modality": "audio", "evidence": "At 0:08, a brief gasp sound can be heard.", "inference": "after hearing the baby's crying, other person or people gasp."}, {"modality": "vision", "evidence": "At 0:07, the baby in yellow is crying.", "inference": "the crying is from the baby in yellow."}, {"modality": "vision", "evidence": "At 0: 09, the father pig opened his mouth in amazement.", "inference": "The baby's crying causes the father to gasp."}, {"modality": "vision", "evidence": "At 0:10, the mother pig who holds baby opens her mouth in amazement, too.", "inference": "The baby's crying causes parents to gasp."}]}
{"id": "video_554__q0001", "video_id": "video_554", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_554_t5s.mp4"], "extra_info": {"question": "Which best characterizes the baby's vocalization as heard and described in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. Forte, indicating a loud and strong sound.", "B. Pianissimo, indicating a very soft and quiet sound.", "C. Legato, indicating a smooth and connected sound.", "D. Staccato, indicating short and detached sounds."], "correct_option": "A"}, "answer_text": "Forte, indicating a loud and strong sound.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:07, the baby is shown in the scene.", "inference": "Identify the baby."}, {"modality": "audio", "evidence": "At 0:07, the baby cries loudly.", "inference": "So the answer is 'Forte, indicating a loud and strong sound.'."}]}
{"id": "video_554__q0002", "video_id": "video_554", "qid": 2, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_554_t5s.mp4"], "extra_info": {"question": "What would happen if there's no pretence that children should pretend to be his baby?"}, "reward_model_ground_truth": {"multi_choice": ["A. There will be two crying children", "B. There will be one crying child", "C. There will be zero crying children", "D. There will be someone to keep his cushion"], "correct_option": "B"}, "answer_text": "There will be one crying children.", "reasoning_steps": [{"modality": "audio", "evidence": "A character says to the baby, telling the baby to pretend to be his child."}, {"modality": "vision", "evidence": "At 0:07,  a baby wearing yellow shirts is crying.", "inference": "The child is crying because he was asked to pretend to be someone else's child."}, {"modality": "audio", "evidence": "At 0:12, another baby's crying voice can be heard.", "inference": "there is another baby is crying not caused by the character's words."}]}
{"id": "video_554__q0003", "video_id": "video_554", "qid": 3, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_554_t5s.mp4"], "extra_info": {"question": "What is Peppa Pig's final decision regarding her 'baby' game and the reason behind it?"}, "reward_model_ground_truth": {"multi_choice": ["A. She decides to keep her cushion as a pretend baby because she finds real babies too noisy.", "B. She decides George can be her baby because he cries realistically and is fun to play with.", "C. She decides to stop playing with babies altogether because George doesn't want to participate.", "D. She decides to ask her parents for a real baby after seeing how well George can imitate one."], "correct_option": "A"}, "answer_text": " She decides to keep her cushion as a pretend baby because she finds real babies too noisy.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:14, Peppa says 'Real babies are quite noisy.'.", "inference": "It indicates that the reason Peppa does not want a real baby anymore is that real babies are too noisy."}, {"modality": "audio", "evidence": "At 0:17, Peppa says 'I'll keep my yellow cushion instead'.", "inference": "Peppa finally decides to keep her cushion as a pretend baby."}, {"modality": "vision", "evidence": "At 0:19, Peppa stuffs her cushion back into her clothe.", "inference": "In the end, Peppa ultimately decide to hold and care for as her pretend baby."}]}
{"id": "video_555__q0000", "video_id": "video_555", "qid": 0, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_555_t5s.mp4"], "extra_info": {"question": "What was the woman doing while the puppy was licking the cola bottle?"}, "reward_model_ground_truth": {"multi_choice": ["A. The woman is drinking coca cola.", "B. The woman is looking for the dog.", "C. The woman is reading a book.", "D. The woman is playing with the puppy."], "correct_option": "B"}, "answer_text": "The woman is looking for the dog.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:07, the puppy was licking the cola bottle.", "inference": "Identify the timestamp."}, {"modality": "audio", "evidence": "At the same time, the woman said 'How did you get here?'.", "inference": "The woman is looking for the dog."}]}
{"id": "video_555__q0001", "video_id": "video_555", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_555_t5s.mp4"], "extra_info": {"question": "Which is displayed when the dog is barking?"}, "reward_model_ground_truth": {"multi_choice": ["A. Two yellow bottle.", "B. One yellow bottle.", "C. Two red bottle", "D. One red bottle."], "correct_option": "D"}, "answer_text": "One red bottle", "reasoning_steps": [{"modality": "audio", "evidence": "The barking sound can be heard.", "inference": "Locate to the screen when the sound can be heard."}, {"modality": "vision", "evidence": "At that time, a cartoon dog exists in one red bottle's pattern.", "inference": "One red bottle is displayed when the dog is barking."}]}
{"id": "video_555__q0002", "video_id": "video_555", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_555_t5s.mp4"], "extra_info": {"question": "What was being introduced when the first knock was heard?"}, "reward_model_ground_truth": {"multi_choice": ["A. Peanut milk", "B. Happy Mahua snack", "C. Voltaren medicine", "D. Coca-Cola"], "correct_option": "C"}, "answer_text": "Coca-Cola", "reasoning_steps": [{"modality": "audio", "evidence": "The girl says to the dog, 'Why did you run here?'", "infefence": "This is the first knock."}, {"modality": "vision", "evidence": "A young man responds to the girl's smile.", "inference": "This is also the first knock."}, {"modality": "vision", "evidence": "After that, two bottles of Coca-Cola exist on the screen, toasting.", "inference": "The answer is Coca-Cola."}]}
{"id": "video_556__q0000", "video_id": "video_556", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_556_t5s.mp4"], "extra_info": {"question": "When one praised the drink 'good' for the first time, how many people in grey clothes were behind him?"}, "reward_model_ground_truth": {"multi_choice": ["A.0.", "B.1.", "C.2.", "D.3."], "correct_option": "C"}, "answer_text": "2", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:21,the one praised the drink 'good' for the first time.", "inference": "This is the first one."}, {"modality": "vision", "evidence": "At 0:21, there are 2 people in grey clothes were behind him.", "inference": "There are two people in grey behind this person."}]}
{"id": "video_556__q0001", "video_id": "video_556", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_556_t5s.mp4"], "extra_info": {"question": "Whose name is most likely to be 'Henry'?"}, "reward_model_ground_truth": {"multi_choice": ["A.The fan.", "B.The anchor.", "C.The bodyguard on the right.", "D.The bodyguard on the left."], "correct_option": "D"}, "answer_text": "The bodyguard on the left.", "reasoning_steps": [{"modality": "vision", "evidence": "Thee are four main characters in the video.", "inference": "Identify the scope the answer."}, {"modality": "vision", "evidence": "At 0:50, the fan gives a bottle of water to the bodyguard on the left.", "inference": "Mark the point that needs attention."}, {"modality": "audio", "evidence": "At 0:51, the fan says 'It's for you, Henry.'", "inference": "The bodyguard on the left is called 'Henry'."}]}
{"id": "video_556__q0002", "video_id": "video_556", "qid": 2, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_556_t5s.mp4"], "extra_info": {"question": "How does the crowd's interaction with Ishowspeed, specifically regarding the beverages, illustrate their engagement with him?"}, "reward_model_ground_truth": {"multi_choice": ["A. They are presenting him with various local tea drinks for him to taste and share his opinion.", "B. They are encouraging him to purchase specific energy drinks that are popular in the region.", "C. They are providing him with drinks to sustain him during a long and exhausting public appearance.", "D. They are offering him a selection of drinks as a token of appreciation for his visit to the area."], "correct_option": "A"}, "answer_text": "They are presenting him with various local tea drinks for him to taste and share his opinion.", "reasoning_steps": [{"modality": "vision", "evidence": "A person brings out of drinks one by one to Ishowspeed and other people.", "inference": "The fans are presenting him with various local tea drinks."}, {"modality": "audio", "evidence": "The crowd of fans are making different comments on drinks.", "inference": "The fans are sharing opinions."}]}
{"id": "video_556__q0003", "video_id": "video_556", "qid": 3, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_556_t5s.mp4"], "extra_info": {"question": "What is the main activity depicted in the clip?"}, "reward_model_ground_truth": {"multi_choice": ["A. IShowSpeed is trying out different Chinese teas while interacting with a crowd of fans.", "B. IShowSpeed is giving out different types of tea to his fans in China.", "C. IShowSpeed is sampling various Chinese snacks and drinks offered by his admirers.", "D. IShowSpeed is talking about famous basketball players with his fans."], "correct_option": "A"}, "answer_text": " IShowSpeed is trying out different Chinese teas while interacting with a crowd of fans.", "reasoning_steps": [{"modality": "vision", "evidence": "IShowSpeed is in a crowd of fans while trying drinks.", "inference": "He is interacting with a crowd of fans."}, {"modality": "audio", "evidence": "IShowSpeed says he has got a lot of drinks.", "inference": "IShowSpeed is trying out differnt Chinese teas."}]}
{"id": "video_557__q0000", "video_id": "video_557", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_557_t5s.mp4"], "extra_info": {"question": "When the actor said 'shake it up with real orange pulp', what was in the scene?"}, "reward_model_ground_truth": {"multi_choice": ["A. Orange", "B. Apple", "C. Monkey", "D. Tiger"], "correct_option": "A"}, "answer_text": "Orange", "reasoning_steps": [{"modality": "audio", "evidence": "The actor said 'shake it up with real orange pulp' at 0:23.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "There is a orange in the scene.", "inference": "It's orange."}]}
{"id": "video_557__q0001", "video_id": "video_557", "qid": 1, "question_type": "causal reasoning", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_557_t5s.mp4"], "extra_info": {"question": "What is the purpose of a man shaking orange juice?"}, "reward_model_ground_truth": {"multi_choice": ["A.Let the animals dance and make way for the cars.", "B.Avoid the danger of animal attacks.", "C.He wants to drink orange juice.", "D.Cheer up the animals."], "correct_option": "D"}, "answer_text": "Cheer up the animals.", "reasoning_steps": [{"modality": "vision", "evidence": "The scene where the man shaking a carton of orange juice is a zoo.", "inference": "The purpose of the man shaking orange juice is about the animals."}, {"modality": "audio", "evidence": "The background music is relaxing and cheerful.", "inference": "The purpose of the man shaking orange juice is to cheer up."}, {"modality": "vision", "evidence": "After the man shook the orange juice, the animals also began to shake their bodies.", "inference": "The purpose of the man shaking orange juice is to cheer up the animals."}]}
{"id": "video_558__q0000", "video_id": "video_558", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_558_t5s.mp4"], "extra_info": {"question": "How many unicorns appeared with the garden?"}, "reward_model_ground_truth": {"multi_choice": ["A.1.", "B.2.", "C.3.", "D.4."], "correct_option": "B"}, "answer_text": "2.", "reasoning_steps": [{"modality": "audio", "evidence": "'garden' appears at the timestamp 1:16", "inference": "The scene in the garden is around this point in time."}, {"modality": "vision", "evidence": "two unicorn appears at the timestamp 1:17", "inference": "there are two unicorn in the garden"}]}
{"id": "video_558__q0001", "video_id": "video_558", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_558_t5s.mp4"], "extra_info": {"question": "How does Peppa think babies are born?"}, "reward_model_ground_truth": {"multi_choice": ["A.Being manufactured by doctors in factories.", "B.Being found by Pig Daddy on the riverbank.", "C.Sliding down from the rainbow.", "D.Being delivered to home by the mail carrier."], "correct_option": "D"}, "answer_text": "Being delivered to home by the mail carrier.", "reasoning_steps": [{"modality": "audio", "evidence": "Peppa says 'then It was time for Evie to be delivered.'", "inference": "Peppa may thinks that babies are born by delivering to home by the mail carrier."}, {"modality": "vision", "evidence": "Evie is delivered by a mail carrier at timestamp 1:58", "inference": "Confirm that Peppa thinks that babies are born by delivering to home by the mail carrier."}]}
{"id": "video_559__q0000", "video_id": "video_559", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_559_t5s.mp4"], "extra_info": {"question": "What is the primary message the speaker aims to convey about her morning routine?"}, "reward_model_ground_truth": {"multi_choice": ["A. She has recently rediscovered her appreciation for music and incorporated it into her mornings due to her musical family.", "B. Her morning skincare routine is very simple, and she prefers just splashing her face with cold water.", "C. She finds piano practice to be a chore, but still makes time for it before starting her work.", "D. She is showing how to transition from personal activities to work efficiently in her home environment."], "correct_option": "A"}, "answer_text": "She has recently rediscovered her appreciation for music and incorporated it into her mornings due to her musical family.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:13, the speaker says 'Because living with a family where everyone loves to sing...'", "inference": "It indicates that she is in a musical family."}, {"modality": "audio", "evidence": "At 0:18, the speaker says '...has really reignited my appreciation for music. '", "inference": "It indicates that having living in a musical family, her appreciation for music has been rekindled."}, {"modality": "vision", "evidence": "At 0:22, she is playing the piano in the morning.", "inference": "So the answer is 'She has recently rediscovered her appreciation for music and incorporated it into her mornings due to her musical family.'"}]}
{"id": "video_559__q0001", "video_id": "video_559", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_559_t5s.mp4"], "extra_info": {"question": "Why does the girl praise the thing she does after she wash her face?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because she used to take piano lessons when she was a child.", "B.Because it is simple to sing and play the instrument for the girl.", "C.Because the environment she's living has good atmosphere.", "D.Because the family she lives with encourage her with words."], "correct_option": "C"}, "answer_text": "Because the environment she's living has good atmosphere.", "reasoning_steps": [{"modality": "vision", "evidence": "she practises the piano after washing face.", "inference": "she starts to play piano."}, {"modality": "audio", "evidence": "she says living with a family that everyone loves to sing and plays the instrument makes her praise music.", "inference": "the environment she's living has good atmosphere"}]}
{"id": "video_56__q0000", "video_id": "video_56", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_56_t5s.mp4"], "extra_info": {"question": "Which part was the artist painting when he mentioned 'by painting large flower still lifes for over a year'?"}, "reward_model_ground_truth": {"multi_choice": ["A.Nose", "B.Lips", "C.Neck", "D.Hair"], "correct_option": "A"}, "answer_text": "Nose", "reasoning_steps": [{"modality": "audio", "evidence": "At 7:00, the artist said 'by painting large flower still lifes for over a year'", "inference": "This is the moment mentioned in the question, identify the part he was painting."}, {"modality": "vision", "evidence": "At 7:00, the artist was painting a woman.", "inference": "Pay attention to which part of the woman he was painting."}, {"modality": "vision", "evidence": "At 7:00, the artist was painting this woman's nose.", "inference": "So the answer is 'Nose'."}]}
{"id": "video_561__q0000", "video_id": "video_561", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_561_t5s.mp4"], "extra_info": {"question": "What is the function of the background music?"}, "reward_model_ground_truth": {"multi_choice": ["A.Suppress background noise.", "B.Highlight the creator's voice.", "C.Capture the audience's attention.", "D.Make the video more appealing."], "correct_option": "D"}, "answer_text": "Make the video more appealing.", "reasoning_steps": [{"modality": "audio", "evidence": "The background music is soft and lyric.", "inference": "Combining the options can rule out A."}, {"modality": "vision", "evidence": "The entire video is showcasing the process of making handicrafts.", "inference": "Music is used to make the video more appealing."}]}
{"id": "video_562__q0000", "video_id": "video_562", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_562_t5s.mp4"], "extra_info": {"question": "Why does the woman choose to do her hot yoga workout around 12 PM?"}, "reward_model_ground_truth": {"multi_choice": ["A. She is likely a 'night owl' and it's her peak energy time.", "B. The hot yoga studio is only open for classes around noon.", "C. She is an 'early bird' but had other commitments in the morning.", "D. The smoothie she made provides the most energy for an afternoon workout."], "correct_option": "A"}, "answer_text": "She is likely a 'night owl' and it's her peak energy time.", "reasoning_steps": [{"modality": "audio", "evidence": "The narrator explains that the optimal time to exercise is during one's peak energy hours, which depends on their 'chronotype', such as being an 'early bird' or a 'night owl'.", "inference": "The time she choose to do her hot yoga workout is determined by her 'chronotype'."}, {"modality": "vision", "evidence": "The video shows timestamps indicating the woman places her workout in the early afternoon.", "inference": "the early afternoon is her peak energy hours."}, {"modality": "vision", "evidence": "she is not working out in the morning.", "inference": "she is not an 'early bird'."}, {"modality": "audio", "evidence": "she presents 'night owl' as the other main chronotype.", "inference": "she is a 'night owl'."}]}
{"id": "video_562__q0001", "video_id": "video_562", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_562_t5s.mp4"], "extra_info": {"question": "What positive outcome is directly linked to exercising during one's most energetic hours?"}, "reward_model_ground_truth": {"multi_choice": ["A. Enhanced immediate energy levels during workouts, as seen in her active movements.", "B. Greater consistency and adherence to a workout routine, exemplified by her structured activities.", "C. Superior physical performance and muscle development, shown by her vigorous exercises.", "D. Improved mental clarity and reduced stress, evidenced by her calm demeanor."], "correct_option": "B"}, "answer_text": "Greater consistency and adherence to a workout routine, exemplified by her structured activities.", "reasoning_steps": [{"modality": "audio", "evidence": "From 0:37 of this video, the recorder says'because if it feels good and natural to workout at those hours, you're more likely to a consistence workout routine.'", "inference": "Greater consistency and adherence to a workout routine is directly linked to excercising during one's most energetic hours."}, {"modality": "vision", "evidence": "From 0:02 of this video, she do a quick pre-workout smoothie at 12:05pm.", "inference": "She is preparing to workout."}, {"modality": "vision", "evidence": "From 0:24 of this video, she is walking to hot yoga at 12:10pm", "inference": "it's her energetic hours, and she is go to workout."}]}
{"id": "video_562__q0002", "video_id": "video_562", "qid": 2, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_562_t5s.mp4"], "extra_info": {"question": "What is the significance of her pre-workout smoothie?"}, "reward_model_ground_truth": {"multi_choice": ["A. Proper nutrition is essential for peak performance.", "B. Her low afternoon energy requires a smoothie, against the advice.", "C. Her chronotype is a night owl needing afternoon fuel.", "D. The hot yoga class demands more energy than her chronotype allows."], "correct_option": "A"}, "answer_text": "Proper nutrition is essential for peak performance.", "reasoning_steps": [{"modality": "audio", "evidence": "The audio establishes the core principle: one should schedule workouts during their natural peak energy hours, as determined by their chronotype, to improve consistency and effectiveness.", "inference": "she schedule her workout during her natural peak energy hours."}, {"modality": "vision", "evidence": "The video shows the woman making a 'quick pre-workout smoothie' at 12:05 pm.", "inference": "The 'quick pre-workout smoothie' is essential for her workout."}, {"modality": "vision", "evidence": "the vision shows that the recipe of the 'quick pre-workout smoothie' contains protein, fruits like banana, blueberries and vegetable like spinach. It provides specific macronutrients (like carbohydrates and protein) as readily available fuel for muscles during strenuous activity and to aid in post-workout recovery.", "inference": "The 'quick pre-workout smoothie' offers a very rich and balanced supply of nutrients, which is is essential for peak performance."}]}
{"id": "video_564__q0000", "video_id": "video_564", "qid": 0, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_564_t5s.mp4"], "extra_info": {"question": "What is the direct cause of the man ultimately losing his briefcase of money?"}, "reward_model_ground_truth": {"multi_choice": ["A. He gave the man in the convertible the money as a reward for helping him recover the cash box.", "B. The thief gave the cash box to the man in the convertible to create further confusion.", "C. The man in the convertible had conspired with the thief to steal the money from the beginning.", "D. He was distracted by the thief's true identity and intentions, giving the man in the convertible an opportunity to take the money."], "correct_option": "D"}, "answer_text": "He was distracted by the thief's true identity and intentions, giving the man in the convertible an opportunity to take the money.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:40, a man in a black offers the man who is stolen in the suit a ride to chase the thief.", "inference": "This establishes their initial role as helpers."}, {"modality": "audio", "evidence": "At 0:56, the man called the thief 'baby'.", "inference": "This man is distracted from his girlfriend."}, {"modality": "vision", "evidence": "At 1:04, while the man is distracted by others, the people in the convertible wave goodbye.", "inference": "This shows that his distraction allowed his 'helpers' to seize the opportunity and steal the briefcase."}]}
{"id": "video_564__q0001", "video_id": "video_564", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_564_t5s.mp4"], "extra_info": {"question": "What visual detail connects the thief's car to the man in the suit?"}, "reward_model_ground_truth": {"multi_choice": ["A. A banana peel on the car's side.", "B. A 'Karl's Bananas' sign on the roof.", "C. 'Just Married' on the rear window.", "D. A dent from hitting a building."], "correct_option": "C"}, "answer_text": "'Just Married' on the rear window.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:11, the man go across a wedding.", "inference": "Get the wedding."}, {"modality": "audio", "evidence": "At 0:11, the man says to the groom, 'Man, I'm sorry about that.  This should cover the ring,' offering a wad of cash as compensation.", "inference": "This establishes the 'wedding' as the specific event where compensation was offered."}, {"modality": "vision", "evidence": "At 0:33, the thief is seen getting into a light blue car.", "inference": "This car is used for the rest of the chase."}, {"modality": "vision", "evidence": "As it drives away at 0:50, the words 'JUST MARRIED' are clearly visible written on the rear window.", "inference": "'Just Married' on the rear window can connect them."}]}
{"id": "video_564__q0002", "video_id": "video_564", "qid": 2, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_564_t5s.mp4"], "extra_info": {"question": "How many distinct times is a 'Bad Guys' movie or a direct sequel/spin-off visually depicted or verbally mentioned in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A. Two", "B. Three", "C. Four", "D. Five"], "correct_option": "C"}, "answer_text": "Four", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:09, a man's voice says 'You may now present the ring.'", "inference": "A man is having a marriage ceremony."}, {"modality": "vision", "evidence": "At 0:10, the man throws a banana into the bridegroom's hand.", "inference": "The man is a bad guy the first time to disrupt others."}, {"modality": "vision", "evidence": "At 0: 23, when the man asks the black woman where is the person, she points to the wrong way.", "inference": "the second time."}, {"modality": "vision", "evidence": "At 0:30, the man drinks another person's bottle of beverage, not paying for it.", "inference": "The third time."}, {"modality": "vision", "evidence": "At 1:00, the man find out that the bad guy is his girlfriend.", "inference": "the fourth time."}]}
{"id": "video_565__q0000", "video_id": "video_565", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_565_t5s.mp4"], "extra_info": {"question": "What is 'that thread' the woman says 'We are cooked on'?"}, "reward_model_ground_truth": {"multi_choice": ["A. The technological advancement shown by the AI video, demonstrating its capabilities.", "B. The technological progress shown by the AI video, revealing its dominance.", "C. The technological evolution shown by the AI video, proving its influence.", "D. The technological development shown by the AI video, showing its potential."], "correct_option": "C"}, "answer_text": "The technological evolution shown by the AI video, proving its influence.", "reasoning_steps": [{"modality": "audio", "evidence": "The interviewer asks for 'one move with AI.'", "inference": "It dicates 'AI'."}, {"modality": "audio", "evidence": "In her response, the woman says, '... it's over.  We are cooked on that thread.'", "inference": "It dicates people are influenced by 'move with AI'."}, {"modality": "vision", "evidence": "Read the visual text overlay which states, 'This Street Interview Is 100% AI' and 'Google Veo 3 just broke the internet'.", "inference": "The technological evolution shown by the AI video, proving its influence."}]}
{"id": "video_565__q0001", "video_id": "video_565", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_565_t5s.mp4"], "extra_info": {"question": "What fundamental paradox does the video present?"}, "reward_model_ground_truth": {"multi_choice": ["A. A person is having an exaggerated emotional reaction that is disproportionate to a simple street interview question.", "B. A video claims to be AI-generated but features human actors who look and sound too realistic to be artificial.", "C. An AI construct is expressing a seemingly authentic emotional realization about the very AI technology that makes her existence possible.", "D. The AI is attempting to trick the viewer into believing it is human, but its emotional response feels unnatural and robotic."], "correct_option": "C"}, "answer_text": "An AI construct is expressing a seemingly authentic emotional realization about the very AI technology that makes her existence possible.", "reasoning_steps": [{"modality": "audio", "evidence": "Her tone, inflection, and word choice ('wild,' 'it's over,' 'we are cooked') convey genuine shock and a sense of profound, almost existential realization.", "inference": "It is so real."}, {"modality": "vision", "evidence": "Her emotions, and the entire scene are '100% AI' and 'Never Actually Happened'.", "inference": "An AI construct is expressing a seemingly authentic emotional realization about the very AI technology that makes her existence possible."}]}
{"id": "video_566__q0000", "video_id": "video_566", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_566_t5s.mp4"], "extra_info": {"question": "What is true about the alumni association event?"}, "reward_model_ground_truth": {"multi_choice": ["A. Organized by Beijing alumni weeks after the shop reopened.", "B. Organized by Beijing alumni months after the shop reopened.", "C. Organized by local alumni months after the shop reopened.", "D. Organized by Beijing alumni committees months after the shop reopened."], "correct_option": "B"}, "answer_text": "Organized by Beijing alumni months after the shop reopened.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:48, the interviewee mentions that the shop reopened around the end of March after the pandemic lockdown.", "inference": "This shop reopened in March."}, {"modality": "audio", "evidence": "At 1:48, he states that they hosted an alumni event around October.", "inference": "Several months passed."}, {"modality": "vision", "evidence": "At 1:48, the Beijing Alumni Association is on his mobile phone.", "inference": "Organized by Beijing alumni months after the shop reopened."}]}
{"id": "video_566__q0001", "video_id": "video_566", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_566_t5s.mp4"], "extra_info": {"question": "Combining the style of this store,which description is incorrect?"}, "reward_model_ground_truth": {"multi_choice": ["A. Limited to camera-themed gatherings.", "B. Restricted to photography enthusiast meetings.", "C. Confined to photographic art community events.", "D. Limited to photography-themed gatherings."], "correct_option": "D"}, "answer_text": " Limited to photography-themed gatherings.", "reasoning_steps": [{"modality": "vision", "evidence": "The style of the store throughout the entire video is related to photography.", "inference": "Combining the options can rule out A."}, {"modality": "audio", "evidence": "At 2:20, he specifies these measures included hosting alumni association meetings and fan club meetings on weekends.", "inference": "Limited to photography-themed gatherings is incorrect."}]}
{"id": "video_566__q0002", "video_id": "video_566", "qid": 2, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_566_t5s.mp4"], "extra_info": {"question": "What is the origin of the coffee shop's name 'Clicked', and what is the most relevant item to this origin shown?"}, "reward_model_ground_truth": {"multi_choice": ["A. From a camera shutter sound; a Phoenix camera is shown with a photo of a couple kissing.", "B. From a camera shutter sound; a Phenix camera is shown with a photo of a couple embracing.", "C. From a camera shutter sound; a Phenix camera is shown with a photo of a couple kissing.", "D. From camera click sounds; a Phenix camera is shown with a photo of a couple kissing."], "correct_option": "C"}, "answer_text": "From a camera shutter sound; a Phenix camera is shown with a photo of a couple kissing.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:27, the interviewee explains that the shop's name is a reference to the sound of a camera shutter clicking.", "inference": "From a camera shutter sound."}, {"modality": "vision", "evidence": "In the end of the video, an old-fashioned silver camera and a kissing couple are shown.", "inference": "The key point indicates that the most relevant symbol item is that a Phenix camera is shown with a photo of a couple kissing."}]}
{"id": "video_567__q0000", "video_id": "video_567", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_567_t5s.mp4"], "extra_info": {"question": "What event made Mitchell call it his 'toughest fight ever'?"}, "reward_model_ground_truth": {"multi_choice": ["A. He was hit by a powerful kick in the first round.", "B. He was hit by a powerful knee in the first round.", "C. He was hit by a powerful elbow in the first round.", "D. He was hit by a powerful punch in the first round."], "correct_option": "B"}, "answer_text": "He was hit by a powerful knee in the first round.", "reasoning_steps": [{"modality": "audio", "evidence": "Listen to the ring announcer at 00:11 declare 'Bryce... Mitchell' as the winner.", "inference": "Bryce Mitchell is the winner."}, {"modality": "vision", "evidence": "Identify the winner, Bryce Mitchell, as the fighter in the camouflage shorts who is subsequently interviewed.", "inference": "Bryce Mitchell is the person interviewed by the interviewer."}, {"modality": "audio", "evidence": "At 00:24, listen to the interviewer ask Mitchell a question, specifically mentioning, 'You took a big knee in the first round'.", "inference": "Bryce Mitchell was hit by a powerful knee in the first round."}, {"modality": "audio", "evidence": "At 00:35, listen to Mitchell's response, 'that was honestly my hardest fight I think ever, bro'.", "inference": "Being hit by a powerful knee in the first round was called by Mitchell his 'toughest fight ever'."}]}
{"id": "video_568__q0000", "video_id": "video_568", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_568_t5s.mp4"], "extra_info": {"question": "In their series of races, how many times did the person who said 'what I say' win?"}, "reward_model_ground_truth": {"multi_choice": ["A.1", "B.2", "C.3", "D.4"], "correct_option": "D"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "The man in white said 'what I say' at 0:12..", "inference": "Identify the person wearing white pants."}, {"modality": "vision", "evidence": "The person wearing white pants runs a head at 0:10.", "inference": "The man won the first time."}, {"modality": "vision", "evidence": "The person wearing white pants runs a head at 0:20.", "inference": "The man won the twice."}, {"modality": "vision", "evidence": "The person wearing white pants runs a head at 0:27.", "inference": "The man won the third times."}, {"modality": "vision", "evidence": "At 0: 42, the person wearing white pants is running a head.", "inference": "The man won the fourth times."}]}
{"id": "video_568__q0001", "video_id": "video_568", "qid": 1, "question_type": "causal reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_568_t5s.mp4"], "extra_info": {"question": "Why did the person do the 'siu' audio?"}, "reward_model_ground_truth": {"multi_choice": ["A.He was taunting his opponent.", "B.He was a Cristiano Ronaldo fan.", "C.He won the race.", "D.He broke a record."], "correct_option": "C"}, "answer_text": "He won the race.", "reasoning_steps": [{"modality": "audio", "evidence": "The person wearing white pants do the 'siu' audio at 0:15", "inference": "Identify the person wearing white pants."}, {"modality": "vision", "evidence": "The person wearing white pants runs a head at 0:10", "inference": "The man in white pants win this race."}]}
{"id": "video_569__q0000", "video_id": "video_569", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_569_t5s.mp4"], "extra_info": {"question": "Which soccer ball in the video is made of cake?"}, "reward_model_ground_truth": {"multi_choice": ["A.The first Ball ", "B.The 3rd ball", "C.'Messi' ball", "D.'C Ronaldo' ball"], "correct_option": "C"}, "answer_text": "'C Ronaldo' ball", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:01, A man scoops cream from one footballs.", "inference": "There's one football made of cake."}, {"modality": "audio", "evidence": "The man says 'don't kick the cake football'", "inference": "There's a football made of cake and the game player shouldn't kick the football."}, {"modality": "audio", "evidence": "At 00:20, the man says 'Subscribe I should kick C Ronaldo's'", "inference": "'Subscribe' action connects with C Ronaldo."}, {"modality": "vision", "evidence": "At 00:20, the left football becomes C Ronaldo' face with a 'Subscribe' icon on it.", "inference": "The left football comes from C Ronaldo."}, {"modality": "audio", "evidence": "At 00:22, the man says 'like I should take Messi's'.", "inference": "'Like' connects with Messi's ball."}, {"modality": "vision", "evidence": "At 00:22, the right football becomes a man's face with a 'like' icon on it.", "inference": "The right football comes from Messi."}, {"modality": "vision", "evidence": "The man turned around to kick the left football.", "inference": "The man choose C Ronaldo' ball."}, {"modality": "vision", "evidence": "At 00:29, the football kicked by the man  breaks up.", "inference": "C Ronaldo' ball is made of cake."}]}
{"id": "video_569__q0001", "video_id": "video_569", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_569_t5s.mp4"], "extra_info": {"question": "The man who yelled 'chicken power' in the video kicked which of the originally placed balls?"}, "reward_model_ground_truth": {"multi_choice": ["A.2", "B.3", "C.4", "D.5"], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision", "evidence": "There are 5 footballs.", "inference": "There are a total of five balls from left to right."}, {"modality": "vision", "evidence": "At 00:04,  the fifth ball from left to right was kicked .", "inference": "The remaining balls were originally the  first to fourth ball counted from left to right"}, {"modality": "vision", "evidence": "At 00:10, the first ball among the remaining balls was kicked by a man.", "inference": "Remaining balls were 2nd, 3rd and 4th balls originally from the left to the right."}, {"modality": "audio", "evidence": "A man says 'Chicken power'", "inference": "This man saying chicken power is going to kick a ball."}, {"modality": "vision", "evidence": "At 00:17, the 3rd ball from left to the right is kicked by this man.", "inference": "The originally placed 4th football from left to right is kicked by the man saying 'chicken power'."}]}
{"id": "video_57__q0000", "video_id": "video_57", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_57_t5s.mp4"], "extra_info": {"question": "What is the tone of young people's opinions about summer?"}, "reward_model_ground_truth": {"multi_choice": ["A.an indifferent tone", "B.a cherished tone", "C.a regretful tone", "D.an excited tone"], "correct_option": "B"}, "answer_text": "a cherished tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 6:32, the word 'SUMMER' appears in the scene.", "inference": "This indicates that they will talk about what they do in summer afterwards."}, {"modality": "audio", "evidence": "At 6:46, the man talks about bees.", "inference": "He may make comments on bees afterwards."}, {"modality": "audio", "evidence": "At 6:48, the man said 'Without them, I don't know what we would do.'.", "inference": "The man speaks of bees in a tone of cherish. So the answer is 'a cherished tone'."}]}
{"id": "video_57__q0001", "video_id": "video_57", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_57_t5s.mp4"], "extra_info": {"question": "What is the tone of the young man and the elderly when introducing things during the harvest season? "}, "reward_model_ground_truth": {"multi_choice": ["A.Excited tone, Tired tone", "B.Gentle/Peaceful tone, Tired tone", "C.Excited tone, Nervous tone ", "D.Gentle/Peaceful tone, Gentle/Peaceful tone"], "correct_option": "C"}, "answer_text": "Excited tone, Nervous tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 8:51, the word 'FALL' appears in the scene.", "inference": "Fall is the season of harvest."}, {"modality": "audio", "evidence": "At 8:51, the young man says'For me, harvest season is probably my favorite ttime...'.", "inference": "So when the young man  introduces the harvest season, he speaks in an excited tone."}, {"modality": "audio", "evidence": "At 9:28, the elderly says 'Harvest for fresh route is a little more intense...'.", "inference": "So the elderly man thinks that harvest season is intense."}, {"modality": "audio", "evidence": "From 9:59, the elderly says 'And so we know we've got about an eight week window to get as much done as we can...'.", "inference": "This indicates that hey have a short time for harvesting, so the elderly man speaks in a nervous tone. So the answer is 'Excited tone, Nervous tone'."}]}
{"id": "video_570__q0000", "video_id": "video_570", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_570_t5s.mp4"], "extra_info": {"question": "What was the primary reason the person decided to eat a second banana?"}, "reward_model_ground_truth": {"multi_choice": ["A.He liked its rich energy", "B.He liked its soft texture", "C.He found its sweet", "D.He found it delicious"], "correct_option": "D"}, "answer_text": "He found it delicious", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:10, the person eats one banana.", "inference": "Mark the point before he gets another banana."}, {"modality": "audio", "evidence": "At 0:12, he comments on the banana in many aspects.", "inference": "Identify that the conclusion will come out."}]}
{"id": "video_570__q0001", "video_id": "video_570", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_570_t5s.mp4"], "extra_info": {"question": "What was the first step the person in the vlog took to combine his meal before eating?"}, "reward_model_ground_truth": {"multi_choice": ["A.He put the corn and potatoes in a new black bowl", "B.He poured the potatoes into the paper cup of corn soup", "C.He ate the corn soup and potatoes from their two separate containers", "D.He poured the corn soup into the bowl of potatoes"], "correct_option": "D"}, "answer_text": "He poured the corn soup into the bowl of potatoes", "reasoning_steps": [{"modality": "vision", "evidence": "The person has a black bowl containing potatoes and a separate paper cup containing corn soup at 0:04.", "inference": "Two separate food components are presented."}, {"modality": "audio", "evidence": "The narrator explicitly states he is going to 'combine' the potatoes and the corn soup at 0:05.", "inference": "Mark the point that needs attention about how to mix the corn soup and potatoes together into one dish."}, {"modality": "vision", "evidence": "He pours the corn soup from the cup directly into the black bowl on top of the potatoes at 0:06.", "inference": "The combining is achieved by transferring the soup into the bowl containing the potatoes."}]}
{"id": "video_573__q0000", "video_id": "video_573", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_573_t5s.mp4"], "extra_info": {"question": "What immediately changed Maverick's mind after he initially refused the mission?"}, "reward_model_ground_truth": {"multi_choice": ["A: The loss of his fighter aircraft", "B: The loss of his height-enhancing insoles", "C: The loss of his military rank", "D: The loss of his substantial salary"], "correct_option": "B"}, "answer_text": "The loss of his height-enhancing insoles", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:22, the superior assigns a mission to Maverick.", "inference": "Mark the beginning point of the question."}, {"modality": "audio", "evidence": "At 0:25, Maverick's expression turns negative about the assignment.", "inference": "Identify Maverick's initial attitude on the mission."}, {"modality": "audio", "evidence": "At 0:28, the superior says that they will take back Maverick's 'Navy-issued height-increasing insoles,' making him the same height as everyone else.", "inference": "That's the threat to Maverick."}, {"modality": "vision", "evidence": "At 0:32, Maverick takes the mission file.", "inference": "Mark Mavorick's change."}, {"modality": "audio", "evidence": "At 0:32, Maverick quickly apologizes and accepts the mission.", "inference": "Maverick changes his initial decision finally."}]}
{"id": "video_574__q0000", "video_id": "video_574", "qid": 0, "question_type": "background&music understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_574_t5s.mp4"], "extra_info": {"question": "How does the musician create a deep note?"}, "reward_model_ground_truth": {"multi_choice": ["A: By pitching down the tapping sound.", "B: By pitching up the tapping sound.", "C: By pitching down the tapping sound and adding a bass sound.", "D: By pitching up the tapping sound and adding a bass sound."], "correct_option": "A"}, "answer_text": "By pitching down the tapping sound.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:20, the musician tap the cup lightly.", "inference": "The start of the sound."}, {"modality": "audio", "evidence": "At 0:45, the musician is explaining that he pitches down the tapping sound.", "inference": "Get the deep note."}, {"modality": "audio", "evidence": "At 0:44, there is a deep note.", "inference": "By pitching down the tapping sound."}]}
{"id": "video_574__q0001", "video_id": "video_574", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_574_t5s.mp4"], "extra_info": {"question": "What is the last sound the person who taps the cup edits before adding the kick drum loop?"}, "reward_model_ground_truth": {"multi_choice": ["A: The drum beat", "B: The original cup tapping sound", "C: The bass sound", "D: The complete track"], "correct_option": "C"}, "answer_text": " The bass sound", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:20, the man in blue tap the cup lightly.", "inference": "Identify the man in blue."}, {"modality": "audio", "evidence": "At 1:25,the man said 'we need some bass..'.", "inference": "Pay attention the next sound."}, {"modality": "audio", "evidence": "At 1:26, the man adds the kick drum loop.", "inference": "The last souned is the bass sound."}]}
{"id": "video_575__q0000", "video_id": "video_575", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_575_t5s.mp4"], "extra_info": {"question": "When the narrator mentions the initial design, what is the robot doing?"}, "reward_model_ground_truth": {"multi_choice": ["A. It is erasing old text.", "B. It is writing 'Hello World'.", "C. It is drawing a picture with a pen.", "D. It is charging."], "correct_option": "A"}, "answer_text": "It is erasing old text.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:03, the narrator introduces Scrubby's initial design.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "At the same time, the robot is erasing old text from the whiteboard.", "inference": "It is erasing old text."}]}
{"id": "video_575__q0001", "video_id": "video_575", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_575_t5s.mp4"], "extra_info": {"question": "How was Scrubby's first drawing with a blue pen on the blackboard?"}, "reward_model_ground_truth": {"multi_choice": ["A. The handwriting is very poor.", "B. The handwriting is neat and clear.", "C. It drew a picture, but it had seams.", "D. It didn't write anything but moved around."], "correct_option": "A"}, "answer_text": "The handwriting is very poor.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:57, Scrubby attempts to draw with a blue pen on the blackboard.", "inference": "Get the point to go ahead."}, {"modality": "audio", "evidence": "At 0:59, the narrator said it wrote poorly.", "inference": "The handwriting is very poor."}]}
{"id": "video_576__q0000", "video_id": "video_576", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_576_t5s.mp4"], "extra_info": {"question": "Which rocket failure scene is shown when explaining a sticky throttle valve is the cause"}, "reward_model_ground_truth": {"multi_choice": ["A. The rocket explodes due to running out of fuel.", "B. The rocket explodes on the launchpad.", "C. The rocket breaks apart after losing control.", "D. The rocket tips over during landing."], "correct_option": "C"}, "answer_text": "The rocket breaks apart after losing control.", "reasoning_steps": [{"modality": "audio", "evidence": "The narration mentions the failure was due to a ''sticky throttle valve''.", "inference": "A sticky throttle value is the cause."}, {"modality": "vision", "evidence": "The corresponding visual shows a rocket engine misfiring and the rocket tumbling out of control.", "inference": "a rocket tumble out, losing control."}, {"modality": "audio", "evidence": "The narration states the specific reason for the failure.", "inference": "The rocket breaks apart after out of control."}]}
{"id": "video_576__q0001", "video_id": "video_576", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_576_t5s.mp4"], "extra_info": {"question": "What is shown when a reporter says critics' negative opinion?"}, "reward_model_ground_truth": {"multi_choice": ["A. A rocket launch fails and explodes.", "B. Elon Musk is working in his office.", "C. A rocket is successfully landing.", "D. Elon Musk is talking to the reporter."], "correct_option": "C"}, "answer_text": "A rocket is successfully landing.", "reasoning_steps": [{"modality": "audio", "evidence": "The reporter says ''Critics say you can't do this''.", "inference": "The reporter says critics' negative opinion."}, {"modality": "audio", "evidence": "Elon Musk immediately replies ''We've done it''.", "inference": "critics' negative opinion is wrong."}, {"modality": "vision", "evidence": "The camera then shows a rocket successfully landing.", "inference": "A rocket is successfully landing."}]}
{"id": "video_577__q0000", "video_id": "video_577", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_577_t5s.mp4"], "extra_info": {"question": "Why does she says 'you stay in there'?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because the baby is ready to be born, but she is not.", "B.Because she doesn't want her child to be born on that day.", "C.Because she doesn't want Manny to share his mother too much", "D.Because she doesn't want to destroy the birthday party."], "correct_option": "B"}, "answer_text": "Because she doesn't want her child to be born on that day.", "reasoning_steps": [{"modality": "vision", "evidence": "Everybody prepared a birthday party for Manny.", "inference": "Today is Manny's birthday."}, {"modality": "audio", "evidence": "Gloria says 'It was bad enough that Manny had to share his mother. I couldn't let him share a birthday, too' in time 1:23.", "inference": "Gloria thinks that the birth may destroy the birthday of Manny's'."}, {"modality": "audio", "evidence": "Gloria says 'You stay in there' to her unborn baby in time 1:30.", "inference": "Gloria dose not want to give the birth today."}]}
{"id": "video_577__q0001", "video_id": "video_577", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_577_t5s.mp4"], "extra_info": {"question": "When the little girl says 'surprise', how is her mood?"}, "reward_model_ground_truth": {"multi_choice": ["A.excited.", "B.displeased.", "C.embarrassed.", "D.pleased"], "correct_option": "C"}, "answer_text": "embarrassed.", "reasoning_steps": [{"modality": "audio", "evidence": "The girl says 'surprise' in time 0:49.", "inference": "The little girl shows up."}, {"modality": "vision", "evidence": "Manny's first kiss happened in time 0:35.", "inference": "Manny's first kiss was interrupted by the surprise party."}, {"modality": "vision", "evidence": "The facial expression of the little girl is calm.", "inference": "The little girl may embarrassed."}]}
{"id": "video_578__q0000", "video_id": "video_578", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_578_t5s.mp4"], "extra_info": {"question": "What is Rechal's attitude when saying congratulations"}, "reward_model_ground_truth": {"multi_choice": ["A.cheerful.", "B.proud.", "C.sorry", "D.claptrap."], "correct_option": "D"}, "answer_text": "claptrap", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:20, Monica askes them to give her the pies,but they forgot.", "inference": "Maybe they are very embrassed."}, {"modality": "vision", "evidence": "At 1:23, Phoebe then presents a trophy they won.", "inference": "Maybe they want to pass this topic."}, {"modality": "audio", "evidence": "Rachel exclaims, 'Congratulations!    '", "inference": "They want to use a wide, celebratory smile to comfort Monica."}]}
{"id": "video_579__q0000", "video_id": "video_579", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_579_t5s.mp4"], "extra_info": {"question": "What is the attitude of a girl in red when one says someone has never had the opportunity to receive his admiration"}, "reward_model_ground_truth": {"multi_choice": ["A.appreciate.", "B.speechless.", "C.surprise", "D.disgust."], "correct_option": "B"}, "answer_text": "speechless", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:14, Sheldon says someone has never had the opportunity to receive his admiration.", "inference": "Sheldon does not have admiration toward someone."}, {"modality": "vision", "evidence": "A girl in red appeared at 0:20.", "inference": "Pay attention to the words and deeds of the girl behind."}, {"modality": "audio", "evidence": "She says to a boy 'you are unbelieveable''.", "inference": "She thinks the boy is unbelievable."}]}
{"id": "video_58__q0000", "video_id": "video_58", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_58_t5s.mp4"], "extra_info": {"question": "What is the tone of the two women regarding the sign?"}, "reward_model_ground_truth": {"multi_choice": ["A. The tone of the first woman is more hurried than that of the second.", "B. The tone of the second woman is more hurried than that of the first.", "C. The tone of the first woman is more firm than that of the second.", "D. The tone of the second woman is more firm than that of the first."], "correct_option": "D"}, "answer_text": "The tone of the second woman is more firm than that of the first.", "reasoning_steps": [{"modality": "vision", "evidence": "at 1:05, a man is tearing down the sign.", "inference": "locate this point."}, {"modality": "vision", "evidence": "after the first woman left, the second woman appeared.", "inference": "locate this woman."}, {"modality": "audio", "evidence": "the man said to the latter woman that 'maybe we should back off a little bit'.", "inference": "he thought they should back off."}, {"modality": "audio", "evidence": "at 1:50 the latter woman said 'no, I will not back off'.", "inference": "she is firm."}]}
{"id": "video_58__q0001", "video_id": "video_58", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_58_t5s.mp4"], "extra_info": {"question": "What real attitude does the man hold towards the ideas of the two women about the sign(the former and the latter)?"}, "reward_model_ground_truth": {"multi_choice": ["A. Approve, Disapprove", "B. Disapprove, Disapprove", "C. Disapprove, approve", "D. Approve, Approve"], "correct_option": "C"}, "answer_text": "Disapprove, approve", "reasoning_steps": [{"modality": "vision", "evidence": "at 1:05, a man is tearing down the sign.", "inference": "locate this point."}, {"modality": "audio", "evidence": "at 1:23, the man said to the first woman that 'I don't think that..'.", "inference": "his attitude is disapproving."}, {"modality": "audio", "evidence": "at 1:44, the latter woman said 'this is crazy'.", "inference": "this is her attitude."}, {"modality": "audio", "evidence": "the man said to the latter woman that 'yeah, this is crazy'.", "inference": "his attitude is approving."}]}
{"id": "video_58__q0002", "video_id": "video_58", "qid": 2, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_58_t5s.mp4"], "extra_info": {"question": "When the issue of moving is mentioned, what emotions do the two parties express in the end?"}, "reward_model_ground_truth": {"multi_choice": ["A. The proposing party is serious, while the listening party thinks it is a joke.", "B. The proposing party is serious, and the listening party also thinks it is serious. ", "C. The listening party is serious, while the proposing party thinks it is a joke.", "D. The listening party is joking, and the proposing party also thinks it is a joke."], "correct_option": "A"}, "answer_text": "The proposing party is serious, while the listening party thinks it is a joke.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:40,a woman said' this is not fun when you are moving'.", "inference": "locate this 'moving'."}, {"modality": "vision", "evidence": "the people who are listening are laughing.", "inference": "the listening party thinks it is a joke."}, {"modality": "audio", "evidence": "At 6:50,a man said' there is no misunderstanding, we want you'.", "inference": "the proposing party is serious"}]}
{"id": "video_580__q0000", "video_id": "video_580", "qid": 0, "question_type": "spatial understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_580_t5s.mp4"], "extra_info": {"question": "When an elephant barks, where can we observe where it is sitting"}, "reward_model_ground_truth": {"multi_choice": ["A.Between the fox and the rabbit", "B.Between the mouse and the deer", "C.Between the fox and the deer", "D.Between the rabbit and the mouse"], "correct_option": "D"}, "answer_text": "Between the rabbit and the mouse", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:14, locate the sound of the elephant barking.", "inference": "Get the barking they talked about."}, {"modality": "vision", "evidence": "identify the elephant and the animals sitting on either side of it.", "inference": "Between the rabbit and the mouse."}]}
{"id": "video_580__q0001", "video_id": "video_580", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_580_t5s.mp4"], "extra_info": {"question": "What is written on the animal nameplate on her right hand side of the rabbit when she introduces herself?"}, "reward_model_ground_truth": {"multi_choice": ["A.CLARK", "B.Nick", "C.Judy", "D.Greg"], "correct_option": "B"}, "answer_text": "Nick", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:25, locate the moment the rabbit introduces herself.", "inference": "Get the introduction they talked about."}, {"modality": "vision", "evidence": "look at the nameplate to the rabbit's right and read the name.", "inference": "Nick"}]}
{"id": "video_581__q0000", "video_id": "video_581", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_581_t5s.mp4"], "extra_info": {"question": "What was in the scene when the real culprit of the mudslide was pointed out? "}, "reward_model_ground_truth": {"multi_choice": ["A. Three persons.", "B. Two persons.", "C. One person.", "D. No person."], "correct_option": "A"}, "answer_text": "Three persons.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:25, the dialogue mentions the event of the mudslide.", "inference": "Mark the point about the question that needs attention."}, {"modality": "audio", "evidence": "At 0:32, the dialogue is 'It was humans who did it.'", "inference": "Identify the very point of scene related to the question."}, {"modality": "vision", "evidence": "There are three main persons walking side by side int he scene.", "inference": "The answer is 'three persons'."}]}
{"id": "video_582__q0000", "video_id": "video_582", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_582_t5s.mp4"], "extra_info": {"question": "Why did the shark pass out?"}, "reward_model_ground_truth": {"multi_choice": ["A. Their helicopter flew too high.", "B. Next to them was a 5,000-ton rocket", "C. Below them was frozen space", "D. They had to jump down by themselves."], "correct_option": "D"}, "answer_text": "They had to jump down by themselves.", "reasoning_steps": [{"modality": "vision", "evidence": "shark passed out at timestamp 0:22", "inference": "narrow the range"}, {"modality": "audio", "evidence": "the wolf says 'no stunt crew'", "inference": "they need to jump by themselves"}]}
{"id": "video_582__q0001", "video_id": "video_582", "qid": 1, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_582_t5s.mp4"], "extra_info": {"question": "Which sound could not be recorded by the camera on the helicopter after 1"}, "reward_model_ground_truth": {"multi_choice": ["A. The wolf announced there was no stunt team.", "B. The counting sound of the countdown", "C. The sound of the shark about to faint", "D. The fish's plea to be let go."], "correct_option": "A"}, "answer_text": "", "reasoning_steps": ""}
{"id": "video_582__q0002", "video_id": "video_582", "qid": 2, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_582_t5s.mp4"], "extra_info": {"question": "When shouting 'Are you kidding me?', what ware they doing?"}, "reward_model_ground_truth": {"multi_choice": ["A.Holding the door handle tightly to prevent being blown away", "B.Trying to open the door", "C.Taking shelter behind the door from the hail", "D.Waiting for the next instruction from their superiors"], "correct_option": "B"}, "answer_text": "Trying to open the door", "reasoning_steps": [{"modality": "audio", "evidence": "hear the line 'Are you kidding me?' at timestamp 0:42", "inference": "locate the timestamp question asks."}, {"modality": "vision", "evidence": "characters are pulling something at timestamp 0:42", "inference": "they may try to open the door"}, {"modality": "vision", "evidence": "the texture of 'open' appears", "inference": "confirm that the answer is 'Trying to open the door'"}]}
{"id": "video_582__q0003", "video_id": "video_582", "qid": 3, "question_type": "summarization", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_582_t5s.mp4"], "extra_info": {"question": "What are they going to do when they say 'oh,right'"}, "reward_model_ground_truth": {"multi_choice": ["A. Rotate counterclockwise.", "B. Rotate clockwise.", "C. Press inward with full force.", "D. Pull outward with full force."], "correct_option": "A"}, "answer_text": "Rotate counterclockwise.", "reasoning_steps": [{"modality": "audio", "evidence": "'oh,right' appear at the timestamp 0:48", "inference": "The answer is around this point in time"}, {"modality": "vision", "evidence": "There is 'open' texture on the hatch", "inference": "this is the hatch in the question"}, {"modality": "vision", "evidence": "The open texture is on a counterclockwise arrow paint.", "inference": "The hatch can be opened by rotating counterclockwise."}]}
{"id": "video_582__q0004", "video_id": "video_582", "qid": 4, "question_type": "summarization", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_582_t5s.mp4"], "extra_info": {"question": "How to open the hatch?"}, "reward_model_ground_truth": {"multi_choice": ["A. Rotate counterclockwise.", "B. Rotate clockwise.", "C. Press inward with full force.", "D. Pull outward with full force."], "correct_option": "A"}, "answer_text": "Rotate counterclockwise.", "reasoning_steps": [{"modality": "vision", "evidence": "hatch appear at the timestamp 0:46", "inference": "narrow the range"}, {"modality": "vision", "evidence": "There is 'open' texture on the hatch", "inference": "this is the hatch in the question"}, {"modality": "vision", "evidence": "The open texture is on a counterclockwise arrow paint.", "inference": "The hatch can be opened by rotating counterclockwise."}]}
{"id": "video_583__q0000", "video_id": "video_583", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_583_t5s.mp4"], "extra_info": {"question": "When it comes to 'No, that was the old way', what color is the largest bottle on the shelf?"}, "reward_model_ground_truth": {"multi_choice": ["A.grey", "B.red", "C.brown", "D.yellow"], "correct_option": "D"}, "answer_text": "yellow", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:47, locate the line 'No, that was the old way'.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "from vision, observe the shelf and identify the color of the largest bottle.", "inference": "yellow"}]}
{"id": "video_583__q0001", "video_id": "video_583", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_583_t5s.mp4"], "extra_info": {"question": "When the protagonist was scolded by the chef, how many cooks totally are there in the kitchen?"}, "reward_model_ground_truth": {"multi_choice": ["A.8", "B.9", "C.10", "D.11"], "correct_option": "D"}, "answer_text": "11", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:13,  the protagonist was scolded by the chef.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "When the protagonist is blamed, there are nine chefs behind him, along with the protagonist and the supervisor.", "inference": "11"}]}
{"id": "video_583__q0002", "video_id": "video_583", "qid": 2, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_583_t5s.mp4"], "extra_info": {"question": "How did the glasses-wearing host comment on the dish the girl was holding?"}, "reward_model_ground_truth": {"multi_choice": ["A. Like two grains of rice on a door.", "B. The type loved by wealthy people in New York.", "C. The amount of food is too much.", "D. Not as good as Kings Hawaiian rolls."], "correct_option": "C"}, "answer_text": "The amount of food is too much.", "reasoning_steps": [{"modality": "vision", "evidence": "The glasses-wearing host appeared at 0:20.", "inference": "Pay attention to the after words of him."}, {"modality": "audio", "evidence": "At 0:17,  the host said that 'the woman in the table 35 wants the sausage for her bonds'.", "inference": "the host thinks the amount of food is too much."}]}
{"id": "video_584__q0000", "video_id": "video_584", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_584_t5s.mp4"], "extra_info": {"question": "Who Chee's girlfriend is?"}, "reward_model_ground_truth": {"multi_choice": ["A.Slushi", "B.Fwench Fwy", "C.Sody Pop", "D.Cofi"], "correct_option": "D"}, "answer_text": "Cofi", "reasoning_steps": [{"modality": "audio", "evidence": "Chikn asks Chee ' Wanna hang out?'.", "inference": "Chikn wants Chee to hang out with him."}, {"modality": "audio", "evidence": "Chee answers  Chikn ' Hanging with my girlfriend'.", "inference": "Chee has girlfriend."}, {"modality": "vision", "evidence": "Then Chee's phone screen displays Cofi's profile picture at 1:47.", "inference": "Cofi is Chee's girlfriend."}]}
{"id": "video_584__q0001", "video_id": "video_584", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_584_t5s.mp4"], "extra_info": {"question": "Can we analyze the relationship between the character in the purple suit and Chikn?"}, "reward_model_ground_truth": {"multi_choice": ["A.friend", "B.couple", "C.enemy", "D.stranger"], "correct_option": "C"}, "answer_text": "enemy", "reasoning_steps": [{"modality": "vision", "evidence": "Chikn lay sadly in the shadow of the character in the purple suit at 2:21.", "inference": "Their relationship is not good."}, {"modality": "audio", "evidence": "The character says to Chikn 'What are reality warping powers for?'.", "inference": "The character tempts Chikn to use reality warping powers ."}, {"modality": "audio", "evidence": "The character says to Chikn 'bend things to your will a little'.", "inference": "The character tempts Chikn to  bend things."}, {"modality": "audio", "evidence": "Chikn says to the character 'What are you talking about'.", "inference": "Chikn doesn't want to do as the  character says."}, {"modality": "vision", "evidence": "The character chuckles at 3:48.", "inference": "They are in an adversarial relationship."}]}
{"id": "video_585__q0000", "video_id": "video_585", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_585_t5s.mp4"], "extra_info": {"question": "What is the relationship between the person wearing a pearl necklace and the little boy?"}, "reward_model_ground_truth": {"multi_choice": ["A.Mother and son.", "B.Sister and twin brother.", "C.Grandma and grandson.", "D.Aunt and nephew."], "correct_option": "C"}, "answer_text": "Grandma and grandson", "reasoning_steps": [{"modality": "vision", "evidence": "the person who wears a pearl necklace appeared at 0:10.", "inference": "Identify the person wearing a pearl necklace."}, {"modality": "audio", "evidence": "The little girl calls the person who wears a pearl necklace 'grandma'.", "inference": "The person wearing a pearl necklace is grandma."}, {"modality": "audio", "evidence": "The woman tells the man that they will have another baby.", "inference": "The little boy exists later will be the little girll's brother."}, {"modality": "audio", "evidence": "The man says 'sorry,mom' to the person who wears a pearl necklace.", "inference": "The man is the son of the person who wears a pearl necklace,the litthe boy is grandson."}]}
{"id": "video_585__q0001", "video_id": "video_585", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_585_t5s.mp4"], "extra_info": {"question": "What was happening when snatching the camera?"}, "reward_model_ground_truth": {"multi_choice": ["A. Crying that the second child exceeds family capacity.", "B. Enjoying the best time with the little daughter.", "C. Complaining about holding another complex kid dance.", "D. Making faces at the camera with the son together."], "correct_option": "A"}, "answer_text": "Crying that the second child exceeds family capacity", "reasoning_steps": [{"modality": "audio", "evidence": "'Now give me the camera.' appear at 0:50.", "inference": "The woman is taking photos."}, {"modality": "vision", "evidence": "At that time,a man is crying at 0:52.", "inference": "Identify the crying man."}, {"modality": "audio", "evidence": "the man says 'I don't want a second kid this at least that capacity.' appear at 0:46.", "inference": "The man is crying that the second child exceeds family capacity."}]}
{"id": "video_586__q0000", "video_id": "video_586", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_586_t5s.mp4"], "extra_info": {"question": "What was the reason the bugs were so crazy the next morning?"}, "reward_model_ground_truth": {"multi_choice": ["A. The child missing a tooth fed a lot of food.", "B. The child with little hair fed a lot of food.", "C. Eric fed a lot of food.", "D. Jeff fed a lot of food."], "correct_option": "A"}, "answer_text": "The child missing a tooth fed a lot of food.", "reasoning_steps": [{"modality": "vision", "evidence": "At 2: 23, at night, a character in black opened the door of the sink.", "inference": "Someone fed the worms."}, {"modality": "vision", "evidence": "At 2: 27, in the morning, Jeff went in the kitchen in amazement.", "inference": "Jeff might find something astonishing in the kitchen."}, {"modality": "vision", "evidence": "At 2: 28, many bugs are wriggling on the floor, the number is crazy.", "inference": "bugs are crazy in number."}, {"modality": "audio", "evidence": "the mother said' Don't worry, hunny. The worms get a little crazy when you over feed them.'", "inference": "the worms are crazy."}]}
{"id": "video_586__q0001", "video_id": "video_586", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_586_t5s.mp4"], "extra_info": {"question": "What food did Jeff take out?"}, "reward_model_ground_truth": {"multi_choice": ["A.watermelon", "B.milk", "C.banana", "D.bread"], "correct_option": "D"}, "answer_text": "bread", "reasoning_steps": [{"modality": "audio", "evidence": "The woman said to the square-faced child'oh, Jeff.' at 0:51.", "inference": "Identify the square-faced child Jeff."}, {"modality": "vision", "evidence": "Jeff took out the bread from the fridge at 0:07", "inference": "Jeff took out the bread."}]}
{"id": "video_586__q0002", "video_id": "video_586", "qid": 2, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_586_t5s.mp4"], "extra_info": {"question": "Why did one person take away the food of the other two?"}, "reward_model_ground_truth": {"multi_choice": ["A. To take home and eat.", "B. To help the two with no appetite.", "C. To give to Eric.", "D. To feed the dog."], "correct_option": "C"}, "answer_text": "To give to Eric.", "reasoning_steps": [{"modality": "vision", "evidence": "At 2: 56, the child with a missing tooth asked Jeff about the hotdog.", "inference": "the child with a missing tooth care the hotdog."}, {"modality": "audio,", "evidece": "At 3:03, a person said 'Cool, then I will give it to Eric as a gift.'", "inference": "A person will give the hotdog to Eric as a gift."}, {"modality": "vision", "evidence": "At 3: 15, the child with a missing tooth pours the cheese and meat in a dish into his bag.", "inference": "The child with a missing tooth takes away the food."}]}
{"id": "video_587__q0000", "video_id": "video_587", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_587_t5s.mp4"], "extra_info": {"question": "What did Sumo do after returning home?"}, "reward_model_ground_truth": {"multi_choice": ["A. Picked up that 1099 form.", "B. Continued eating ice cream.", "C. Put the lampshade over head.", "D. Hung the wet paper to dry."], "correct_option": "C"}, "answer_text": "Put the lampshade over head.", "reasoning_steps": [{"modality": "audio", "evidence": "A kid says 'hey mum' to a woman at timestamp 3:37", "inference": "This is their home"}, {"modality": "vision", "evidence": "Sumo put the lampshade over his head.", "inference": "The answer is 'Put the lampshade over head'."}]}
{"id": "video_587__q0001", "video_id": "video_587", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_587_t5s.mp4"], "extra_info": {"question": "What is the third letter of Larry's license plate number?"}, "reward_model_ground_truth": {"multi_choice": ["A.7", "B.8", "C.9", "D.E"], "correct_option": "B"}, "answer_text": "8", "reasoning_steps": [{"modality": "audio", "evidence": "An middle-aged man reply to the woman's greeting at timestamp 3:54", "inference": "The man is Larry"}, {"modality": "vision", "evidence": "Larry gets out from a grey car at timestamp 0:45", "inference": "The car is Larry's"}, {"modality": "vision", "evidence": "Larry's license plate appear at timestamp 0:01 saying '7E89ZY'", "inference": "The third letter is 8."}]}
{"id": "video_588__q0000", "video_id": "video_588", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_588_t5s.mp4"], "extra_info": {"question": "How many people born in the 1990s are there in total?"}, "reward_model_ground_truth": {"multi_choice": ["A.1", "B.2", "C.3", "D.4"], "correct_option": "C"}, "answer_text": "3", "reasoning_steps": [{"modality": "audio", "evidence": "the man asks 'those who born in the 1990s stand up' at the timestamp 0:50", "inference": "need to count the number of person after the man give his order."}, {"modality": "vision", "evidence": "A man in black is marked 1990 at the timestamp 0:50", "inference": "the first one born in the 1990s is found"}, {"modality": "vision", "evidence": "A man in black and white is marked 1992 at the timestamp 0:50", "inference": "the second one born in the 1990s is found"}, {"modality": "vision", "evidence": "A woman  is marked 1997 at the timestamp 0:54", "inference": "the third one born in the 1990s is found"}]}
{"id": "video_588__q0001", "video_id": "video_588", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_588_t5s.mp4"], "extra_info": {"question": "Which was not said by someone wearing ink-colored clothes?"}, "reward_model_ground_truth": {"multi_choice": ["A.Those who think they are good-looking stand up", "B.The actor who played the male lead stand up", "C.Those who have never played the leading woman role stand up", "D.Those in shoes stand up"], "correct_option": "C"}, "answer_text": "Those who have never played the leading woman role stand up", "reasoning_steps": [{"modality": "vision", "evidence": "a man in black and white appears at timestamp 0:28", "inference": "he is the one wearing ink-colored clothes"}, {"modality": "audio", "evidence": "he says 'Those who think they are good-looking stand up' at timestamp 0:01", "inference": "option A is said"}, {"modality": "audio", "evidence": "he says 'The actor who played the male lead stand up' at timestamp 0:32", "inference": "option B is said"}, {"modality": "audio", "evidence": "he says 'Those in shoes stand up' at timestamp 1:07", "inference": "option D is said"}]}
{"id": "video_589__q0000", "video_id": "video_589", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_589_t5s.mp4"], "extra_info": {"question": "Which is the second key node in the change of the protagonist's state of mind?"}, "reward_model_ground_truth": {"multi_choice": ["A. The appearance of the teacher being angry", "B. Eye corrector", "C. Paintings in one's own style", "D. The second painting style required by the teacher"], "correct_option": "D"}, "answer_text": "The second painting style required by the teacher", "reasoning_steps": [{"modality": "audio", "evidence": "The teacher says 'draw more than what you can see' at 4:22.", "inference": "The teacher gives a new instruction."}, {"modality": "vision", "evidence": "Then student draws a realism picture at 4:33.", "inference": "He begins drawing under the new instruction."}, {"modality": "vision", "evidence": "Then student gives up the pior picture and draws another one 4:44.", "inference": "He changes his mind and draws as requested by the teacher."}]}
{"id": "video_59__q0000", "video_id": "video_59", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_59_t5s.mp4"], "extra_info": {"question": "What would the long-haired man do if the guy hadn't sung out? "}, "reward_model_ground_truth": {"multi_choice": ["A. Cheer for him", "B. Feel lost for him", "C. Feel admiration for him", "D. Feel disappointed in him"], "correct_option": "D"}, "answer_text": "Feel disappointed in him.", "reasoning_steps": [{"modality": "audio", "evidence": "At 00:01, the short-haired woman said, 'I love the boy, but he's always pitchy, and he refuses to use his diaphragm.'", "inference": "she thought the boy wouldn't sing out."}, {"modality": "audio", "evidence": "At 00:05, the boy sang out passionately, 'cut my life into pieces.'", "inference": "the boy sang out."}, {"modality": "vision", "evidence": "At 00:05, the long-haired woman were so surprised.", "inference": "the long-haired woman was surprised by the boy."}, {"modality": "vision", "evidence": "At 00:05, the short-haired woman had a look of surprise in the picture.", "inference": "the short-haired woman was surprised by the boy."}, {"modality": "audio", "evidence": "At 00:20, the long-haired woman praised the boy, 'that's going well.'", "inference": "so the long-haired woman would feel disappointed in the boy, if he didn't sang out."}]}
{"id": "video_59__q0001", "video_id": "video_59", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_59_t5s.mp4"], "extra_info": {"question": "The difference in the tone of voice between the man and the child when  playing basketball and speaking. "}, "reward_model_ground_truth": {"multi_choice": ["A. sadder than before", "B. merrier than before", "C. more restrained than before", "D. more excited than before"], "correct_option": "B"}, "answer_text": "merrier than before. ", "reasoning_steps": [{"modality": "vision", "evidence": "At 05:12, the man hold a  basketball.", "inference": "the man was playing basketball."}, {"modality": "audio", "evidence": "At 05:16, the man said, 'Ta-da!'", "inference": "the man was happy."}, {"modality": "audio", "evidence": "At 05:17, the children screamed.", "inference": "when playing basketball and speaking, the man and the children were merrier than before."}]}
{"id": "video_590__q0000", "video_id": "video_590", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_590_t5s.mp4"], "extra_info": {"question": "What is the word in the frame before the guest-related term mentioned by the woman?"}, "reward_model_ground_truth": {"multi_choice": ["A.to wipe", "B.do the dishes", "C.to polish", "D.to load the dishwater"], "correct_option": "A"}, "answer_text": "to wipe", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:13, the woman said 'you can polish silverware before your guests arrive.'.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "At that moment, the vision shows 'to polish'.", "inference": "Pay attention the before."}, {"modality": "vision", "evidence": "Before that moment, 'to wipe'is showed in the frame.", "inference": "to wipe."}]}
{"id": "video_590__q0001", "video_id": "video_590", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_590_t5s.mp4"], "extra_info": {"question": "Which task was introduced after 'to change the bed sheets'?"}, "reward_model_ground_truth": {"multi_choice": ["A.to replace light bulbs.", "B.stock up on (something).", "C.to change the bed sheets", "D.to go grocery shopping."], "correct_option": "A"}, "answer_text": "to replace light bulbs.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:48, the woman said 'to change the bed sheets'.", "inference": "Pay attention to the next action after it."}, {"modality": "vision", "evidence": "At 7:00, there is a phrase 'to replace light bulbs' on the screen.", "inference": "After that, there is 'to replace light bulbs'."}]}
{"id": "video_591__q0000", "video_id": "video_591", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_591_t5s.mp4"], "extra_info": {"question": "What is the direct reason for the chef's fifth scrape the mold?"}, "reward_model_ground_truth": {"multi_choice": ["A.Smooth out the chocolate by scraping", "B.Prevent the chocolate from forming curled edges", "C.Prevent demolding", "D.Allow the chocolate to wrap evenly"], "correct_option": "B"}, "answer_text": "Prevent the chocolate from forming curled edges", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:50, the man scraped the chocolate for the fifth time.", "inference": "The man scraped chocolate for the fifth time."}, {"modality": "audio", "evidence": "At 3:23, the man said that in this way, the chocolate solidified inside the mold, and the side of the mold would not form a straight edge, but actually, there would be a little inclination.", "inference": "Prevent the chocolate from forming curled edges"}]}
{"id": "video_591__q0001", "video_id": "video_591", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_591_t5s.mp4"], "extra_info": {"question": "Why does this chef bake chocolate after it has solidified?"}, "reward_model_ground_truth": {"multi_choice": ["A.To shape the chocolate into the desired shape.", "B.To connect the two pieces of chocolate.", "C.To enhance the texture of the chocolate.", "D.To engrave the shop logo on the chocolate."], "correct_option": "B"}, "answer_text": "To enhance the texture of the chocolate.", "reasoning_steps": [{"modality": "audio", "evidence": "The chef says, 'If you want to, you can put your self logo' (note: corrected 'self' to 'self -' as likely typo for 'own').", "inference": "Suggests an action (adding logo) related to post-solidification steps, but not directly the baking reason yet."}, {"modality": "vision", "evidence": "He hasn't put any logo on his chocolate before.After saying that, 2 chocolates have logos on them.", "inference": "The visual shows logo addition as a new step, but we need to link it to baking. However, since the correct answer is about connecting chocolate pieces, we re-align: The process implies post-solidification actions (like baking) serve a structural purpose (connecting), and the logo is a secondary detail."}]}
{"id": "video_591__q0002", "video_id": "video_591", "qid": 2, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_591_t5s.mp4"], "extra_info": {"question": "When it comes to the chef's favorite part, how many pieces of the second is ingredient on top?"}, "reward_model_ground_truth": {"multi_choice": ["A.2", "B.3", "C.4", "D.5"], "correct_option": "C"}, "answer_text": "4", "reasoning_steps": [{"modality": "aduio", "evidence": "At 6 minutes and 13 seconds, the man said, Put some coconut and crispy beads.", "inference": "The second kind of small material is brittle beads."}, {"modality": "vision", "evidence": "The man put four crisp beads at 6 minutes and 13 seconds.", "inference": "The quantity of the second small material added is 4."}]}
{"id": "video_592__q0000", "video_id": "video_592", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_592_t5s.mp4"], "extra_info": {"question": "Which two numbers correspond to the handmade crafts that are relatively similar for uses and appearance 改?"}, "reward_model_ground_truth": {"multi_choice": ["A.3rd and 9th", "B.2nd and 5th", "C.1st and 7th", "D.4th and 8th"], "correct_option": "A"}, "answer_text": "3rd and 9th", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:22 of the video, the 3rd handmade craft is a miniature mailbox made from a milk carton", "inference": "the 3rd handmade craft is a rectangular box with a slot."}, {"modality": "audio", "evidence": "At  2:43 of the video,the narrator explains that the mailbox's purpose is for exchanging letters with friends.", "inference": "The 3rd handmade craft is made to connect with friends and deepen your bond."}, {"modality": "vision", "evidence": "At 7:47 of the video, the 9th craft is a box resembling a ticket dispenser, filled with love coupons.", "inference": "the 9th handmade craft is also a rectangular box with a slot."}, {"modality": "audio", "evidence": "At 8:04 of the video,the narrator explains that those love coupon make unfogettable moments and get fun together with friend.", "inference": "The 9th handmade craft is also made to deepen your bond with friends."}]}
{"id": "video_592__q0001", "video_id": "video_592", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_592_t5s.mp4"], "extra_info": {"question": "Which of the following is the correct production sequence for the small handmade craft?"}, "reward_model_ground_truth": {"multi_choice": ["A.Shape paper into a bouquet → Roll several → Trim the second paper to wrap the first →  Wrap them with newspaper", "B.Make CD transparent → Cut into small rectangles → Stick photo on transparent square → Draw → Cover with another square → Hang with string", "C.Cut milk carton as in the video → Trim with stickers/scissors → Add favorite elements → Paint →  Fold paper mails into handmade mailbox", "D.Cut flower shape → Stack and stick on paper → Make 3D →  Close the bottom layer"], "correct_option": "C"}, "answer_text": "Cut milk carton as in the video → Trim with stickers/scissors → Add favorite elements → Paint → Fold paper mails into handmade mailbox", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:23 of the video, it shows a milk carton being cut to form the mailbox's basic shape.", "inference": "The first step is cutting milk carton as in the video"}, {"modality": "vision", "evidence": "At 2:28 of the video, the carton is then taped into shape.", "inference": "The second step is trimming with stickers"}, {"modality": "vision", "evidence": "At 2:36 of the video, a mail slot is cut into the front by a pair of scissors.", "inference": "The second step is trimming with scissors."}, {"modality": "vision", "evidence": "At  2:40 of the video, the mailbox is decorated by painting it and drawing details with a marker.", "inference": "Later steps are adding favorite elements and painting."}, {"modality": "audio", "evidence": "At 2:43 of the video,the narrator explains that the mailbox's purpose is for exchanging letters.", "inference": "She will fold paper mails into handmade mailbox after finish the handmade craft."}]}
{"id": "video_593__q0000", "video_id": "video_593", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_593_t5s.mp4"], "extra_info": {"question": "Why was the last sentence of the heavier member in the main group after he passed all the regions?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is really fond of playing.", "B. For some reason, he doesn't want to leave the protagonist playing alone.", "C. He is unwilling to abandon the protagonist.", "D. He is jealous of the protagonist."], "correct_option": "C"}, "answer_text": "He is unwilling to abandon the protagonist.", "reasoning_steps": [{"modality": "vision", "evidence": "At 4:50, the bigger orange cat endures fire and dangerous enviroment to come to the little orange cat.", "inference": "This establishes a high-stakes, life-threatening context for their actions."}, {"modality": "audio", "evidence": "At 4:54, the heavier member states, 'That was too close, even for me.'", "inference": "This confirms the danger was significant."}, {"modality": "audio", "evidence": "At 05:02, he responds with the line, 'I couldn't let you have all the fun.'", "inference": "The bigger member says a joke as a relief to his teammate that was not left behind in a deadly environment."}]}
{"id": "video_593__q0001", "video_id": "video_593", "qid": 1, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_593_t5s.mp4"], "extra_info": {"question": "What would happen if the protagonist hadn't been caught by another partner with a rope?"}, "reward_model_ground_truth": {"multi_choice": ["A. The mission ends in failure.", "B. Remaining companions will finish the mission first, then rescue him", "C. Remaining companions will rescue him first, then finish the mission.", "D. The mission doesn't fail; remaining companions can keep going.."], "correct_option": "D"}, "answer_text": "The mission doesn't fail; remaining companions can keep going.", "reasoning_steps": [{"modality": "vision", "evidence": "At 3:13, one team member is captured.", "inference": "He is in danger."}, {"modality": "audio", "evidence": "At 3:19, the team leader immediately orders the others to 'carry on with the mission'.", "inference": "The leader sticks to the mission to achieve their objective."}, {"modality": "vision", "evidence": "At 04:11, the protagonist nearly falls into a fondue vat.", "inference": "the event central to the hypothetical question."}]}
{"id": "video_593__q0002", "video_id": "video_593", "qid": 2, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_593_t5s.mp4"], "extra_info": {"question": "Why doesn't the big orange cat go to the loading dock?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because he found the cheese production line more interesting.", "B. Because he didn't need to get that truck.", "C. Because he thought the little orange cat was in danger.", "D. Because he found that the loading and unloading dock was not the correct route."], "correct_option": "C"}, "answer_text": "Because he thought the little orange cat was in danger.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:32, the Ox says 'You are in the pine cone.'", "inference": "They had arrived at 'pine cone'."}, {"modality": "audio", "evidence": "At 2:44, the little orange cat says 'Look, look, here is the loading dock.'", "inference": "They found the loading dock."}, {"modality": "vision", "evidence": "At 3:20, the little orange cat was taken to the cheese production line.", "inference": "The little orange cat was in danger."}, {"modality": "audio", "evidence": "At 3:26, the Ox says 'get to the load dock before the workers return'.", "inference": "Ox asks the big orange cat to carry on mission."}, {"modality": "vision", "evidence": "At 4:44, the big orange cat jumped from the handrail to check the little orange cat.", "inference": "The big orange cat didn't get to the loading dock."}]}
{"id": "video_594__q0000", "video_id": "video_594", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_594_t5s.mp4"], "extra_info": {"question": "What problem shown in the video does it solve?"}, "reward_model_ground_truth": {"multi_choice": ["A. Developed after Facebook/Twitter to use social networks for food sharing.", "B. Aimed to solve farm food waste, the first stage in the visual chart.", "C. Solves a problem with global environmental impact second only to the US and China.", "D. Solves a problem causing a 4-person household to waste ~$1,500 in food yearly."], "correct_option": "C"}, "answer_text": "Solves a problem with global environmental impact second only to the US and China.", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:02, the narrator states the FoodCam was invented in 1999.", "inference": "Get the invention time."}, {"modality": "vision", "evidence": "A timeline at 1:03 confirms this, showing the FoodCam icon in 1999.", "inference": "The FoodCan's appearing time is really early."}, {"modality": "audio", "evidence": "The narrator explicitly says this was 'before social media as we know it.'", "inference": "This establishes the time frame of the invention."}, {"modality": "audio", "evidence": "Later in the video, he states, 'If global food waste were a country, it would be the third-largest emitter of greenhouse gases, just behind China and the United States' from2:08.", "inference": "The narrator explains the severity of the food waste problem."}, {"modality": "vision", "evidence": "A corresponding visual graphic at 2:14 displays a ranking: 1. China flag, 2. USA flag, 3. An icon representing food waste.", "inference": "The scale of the food waste problem is so immense that its environmental impact is now understood to be surpassed only by China and the United States."}]}
{"id": "video_595__q0000", "video_id": "video_595", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_595_t5s.mp4"], "extra_info": {"question": "Which option accurately describes the hunting clothes recovered by the divers?"}, "reward_model_ground_truth": {"multi_choice": ["A.Black and almost torn hunting clothes", "B.Black hunting clothes with an ID inside", "C.Green and almost torn hunting clothes with an ID inside", "D.None of the above"], "correct_option": "D"}, "answer_text": "None of the above", "reasoning_steps": [{"modality": "vision", "evidence": "The hunting clothes displays at 6:27.", "inference": "It is green and didn't been torn."}, {"modality": "audio", "evidence": "Mike said Mike's friend pulled out a hunting license from the pocket of this cloth.", "inference": "There is an ID inside the hunting clothes."}]}
{"id": "video_596__q0000", "video_id": "video_596", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_596_t5s.mp4"], "extra_info": {"question": "When mentioning that 'she' was unaware of the truth, what was the woman wearing?"}, "reward_model_ground_truth": {"multi_choice": ["A: A grey top", "B: A white top", "C: A red dress", "D: A black dress"], "correct_option": "A"}, "answer_text": "A grey top", "reasoning_steps": [{"modality": "audio", "evidence": "The man says he keeps playing the role as usual, like normal husband at 2:08.", "inference": "His wife doesn't know the truth."}, {"modality": "vision", "evidence": "At 2:09, the man and woman stay at the sofa.", "inference": "locate the point of the answer"}, {"modality": "vision", "evidence": "At 2:09, her upper cloth is grey", "inference": "The woman wears a grey top."}]}
{"id": "video_597__q0000", "video_id": "video_597", "qid": 0, "question_type": "background&music understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_597_t5s.mp4"], "extra_info": {"question": "When the solar water pump system was shown first, what sound could be heard?"}, "reward_model_ground_truth": {"multi_choice": ["A: The sound of running water", "B: Bird chirping", "C: Josephine's voice", "D: A buzzing sound"], "correct_option": "D"}, "answer_text": " A buzzing sound", "reasoning_steps": [{"modality": "vision", "evidence": "At 4:36, the solar panel pump system is shown first on the screen.", "inference": "Set the point to check the sound information."}, {"modality": "audio", "evidence": "At the same moment, a buzzing sound is heard clearly.", "inference": "A buzzing sound could be heard when the solar water pump system was shown first."}]}
{"id": "video_597__q0001", "video_id": "video_597", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_597_t5s.mp4"], "extra_info": {"question": "When Josephine talks about having to carry water away, what does she point to?"}, "reward_model_ground_truth": {"multi_choice": ["A: The ground beneath her feet", "B: The water pipe next to her", "C: The solar panels in the distance", "D: The trees behind her"], "correct_option": "A"}, "answer_text": "The ground beneath her feet", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:41, Josephine said that her biggest challenge was water.", "inference": "Pay attention to the related content when talking about water next."}, {"modality": "audio", "evidence": "At 1:50, Josephine says talks she had to carry water from a river 400 meters away from her home every day  .", "inference": "Josephine talks about having to carry water away."}, {"modality": "vision", "evidence": "At the 2:09, Josephine's hand points to the ground beneath her feet.", "inference": "she point to the ground beneath her feet ."}]}
{"id": "video_598__q0000", "video_id": "video_598", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_598_t5s.mp4"], "extra_info": {"question": "What was the barber's reaction when Julian commented on his own haircut?"}, "reward_model_ground_truth": {"multi_choice": ["A: He advised not to let hair grow so long and satisfied .", "B: He said the new haircut suited well and confident.", "C: He asked to pay for this haircut and was satisfied with the result.", "D: He invited to come back for haircuts and confident."], "correct_option": "A"}, "answer_text": "He advised not to let hair grow so long .", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:16, Julian comments on the new haircut 'I love it'.", "inference": "Locate his reaction"}, {"modality": "audio", "evidence": "At 4:22, the barber says to him, 'don't let your hair get so long next time.'", "inference": "He advised not to let hair grow so long."}, {"modality": "vision", "evidence": "At 4:20, The barber readily took Julian's money", "inference": "he was very satisfied with his own haircut work."}]}
{"id": "video_599__q0000", "video_id": "video_599", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_599_t5s.mp4"], "extra_info": {"question": "Why does a relatively large number appear?"}, "reward_model_ground_truth": {"multi_choice": ["A. Indicates how many minutes the noodles need to be cooked.", "B. Indicates how many years he has lived in America.", "C. Indicates how many years he has been eating the noodles", "D. Indicates how many seasoning packets the noodles need"], "correct_option": "C"}, "answer_text": "Indicates how many years he has been eating the noodles", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:43, people ask the chef how many years you've been eating it.", "inference": "Bring about the question."}, {"modality": "audio", "evidence": "At 0:44, Chef responds 20.", "inference": "Chef loves eating it over 20 years."}, {"modality": "vision", "evidence": "At 0:44, It shows the number 20.", "inference": "It indicates how many years he has been eating the noodles: 20 years."}]}
{"id": "video_599__q0001", "video_id": "video_599", "qid": 1, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_599_t5s.mp4"], "extra_info": {"question": "What would be the most likely to happen if he was at home"}, "reward_model_ground_truth": {"multi_choice": ["A. Eat a pack of instant noodles with chopsticks.", "B. Eat two packs of instant noodles with chopsticks.", "C. Eat a pack of instant noodles with a fork.", "D. Eat two packs of instant noodles with a fork."], "correct_option": "B"}, "answer_text": "Eat two packs of instant noodles with chopsticks.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:58,It can be seen that the chef is cooking instant noodles, so the theme this time is cooking instant noodles. ", "inference": "question will be raised regarding his possible habits at home."}, {"modality": "audio", "evidence": "At 5:52, the chef mentions usually he eats instant noodles with chopsticks.", "inference": "He prefers using chopsticks for eating instant noodles."}, {"modality": "audio", "evidence": "At 0:30, the chef mentions he mix 2 packs of instant noodles at home when he's hungry.", "inference": "He will eat 2 packs of instant noodles."}]}
{"id": "video_599__q0002", "video_id": "video_599", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_599_t5s.mp4"], "extra_info": {"question": "What is the red face emoji used for?"}, "reward_model_ground_truth": {"multi_choice": ["A. Seasoning packet.", "B. Dried vegetables.", "C. Instant noodles.", "D. Additional oil."], "correct_option": "A"}, "answer_text": "Fry one side", "reasoning_steps": [{"modality": "audio", "evidence": "locate the chef's comment 'this egg yolk is creamy"}, {"modality": "vision", "evidence": "review the footage prior to the comment to see how the egg was cooked"}]}
{"id": "video_599__q0003", "video_id": "video_599", "qid": 3, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_599_t5s.mp4"], "extra_info": {"question": "What did the chef consider perfect right before adding the oil? "}, "reward_model_ground_truth": {"multi_choice": ["A. Noodles absorb all the moisture.", "B. Mix the two pieces of dough and put in.", "C. Add a perfectly fried egg.", "D. The spicy and sweet taste is delicious."], "correct_option": "A"}, "answer_text": "Noodles absorb all the moisture.", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:50, chef says 'all water become to dry. Like this it's perfect.'", "inference": "The perfect time to add oil is when the water becomes dry."}, {"modality": "vision", "evidence": "At 3:55, instant noodles absorb water.", "inference": "Before adding the oil, chef have to wait noodles absorb all the moisture."}]}
{"id": "video_6__q0000", "video_id": "video_6", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_6_t5s.mp4"], "extra_info": {"question": "Which is not mentioned between the first and second occurrence of the same place name in the text?"}, "reward_model_ground_truth": {"multi_choice": ["A. Safety is important.", "B. Cell phones are on sky.", "C. You are in movie.", "D. No door."], "correct_option": "C"}, "answer_text": "You are in movie.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:00, the text of 'Alasa' appears at the first time.", "inference": "It's the starting point of the examined time period."}, {"modality": "vision", "evidence": "At 0:25, the text of 'Alasa' appears at the second time.", "inference": "It's the ending point of the examined time period."}, {"modality": "audio", "evidence": "There are 5 segments of character dialogue during this period.", "inference": "During the time, 'You are in movie' is not mentioned."}]}
{"id": "video_60__q0000", "video_id": "video_60", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_60_t5s.mp4"], "extra_info": {"question": "What was the woman's tone of voice before and after (she found out about) the pregnancy?"}, "reward_model_ground_truth": {"multi_choice": ["A. Approve->objective", "B. Objective->things went against expectations.", "C. Approving->things went against expectations.", "D. Objective->argumentative."], "correct_option": "C"}, "answer_text": "Approving->things went against expectations.", "reasoning_steps": [{"modality": "vision", "evidence": "at 8:07,a pregnant woman is happy.", "inference": "locate this point."}, {"modality": "audio", "evidence": "after a man said' she is happy, isn't she?', a woman said 'yeah'.", "inference": "she is approving."}, {"modality": "audio", "evidence": "then, she said 'when things went against expectations....'.", "inference": "things went against expectations."}]}
{"id": "video_60__q0001", "video_id": "video_60", "qid": 1, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_60_t5s.mp4"], "extra_info": {"question": "What would happen if the man hadn't gone back to play the piano?"}, "reward_model_ground_truth": {"multi_choice": ["A. He would lose his dignity in front of the child", "B. Nothing would happen.", "C. He would gain the respect of the audience present.", "D. It would cause the pregnant woman to miscarry."], "correct_option": "C"}, "answer_text": "He would gain the respect of the audience present.", "reasoning_steps": [{"modality": "vision", "evidence": "a man is playing the piano.", "inference": "locate this point."}, {"modality": "vision", "evidence": "At 6:44, a pregnant woman will give birth to a baby.", "inference": "maybe she is trying hard to sit there."}, {"modality": "audio", "evidence": "the pregnant woman said 'dad is next'.", "inference": "maybe the next man is her dad."}, {"modality": "vision", "evidence": "a man who is playing piano is her dad.", "inference": "get the person."}, {"modality": "vision", "evidence": "her dad will go when she has to go to hospital.", "inference": "her dad would gain the respect of the audience present if the man hadn't gone back to play the piano."}]}
{"id": "video_60__q0002", "video_id": "video_60", "qid": 2, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_60_t5s.mp4"], "extra_info": {"question": "What real emotion does the man reveal when he talks about his birthday?"}, "reward_model_ground_truth": {"multi_choice": ["A. Happy", "B. Proud", "C. Disappointed", "D. Unhappy"], "correct_option": "D"}, "answer_text": " Unhappy", "reasoning_steps": [{"modality": "vision", "evidence": "At 8:17, A man is sitting on the sofa and talking about something.", "inference": "locate this point."}, {"modality": "audio", "evidence": "At 8:17, a man said ' a birthday..'", "inference": "locate this point again."}, {"modality": "audio", "evidence": "he said 'no expectations'.", "inference": "he is unhappy."}]}
{"id": "video_60__q0003", "video_id": "video_60", "qid": 3, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_60_t5s.mp4"], "extra_info": {"question": "What events occurred between the man riding his bike and the moment he called for help?"}, "reward_model_ground_truth": {"multi_choice": ["A. Argued over grapes.", "B. A family sang songs together.", "C. Tried to make the bacon fly.", "D. The girls helped each other style their hair."], "correct_option": "B"}, "answer_text": "A family sang songs together.", "reasoning_steps": [{"modality": "vision", "evidence": "at 5:24, a man is riding a bike.", "inference": "locate it."}, {"modality": "audio", "evidence": "at 11:00, a man is asking for a help.", "inference": "locate it."}, {"modality": "vision", "evidence": "at 6:11, a family sang songs together.", "inference": "it happened between the man riding his bike and the moment he called for help."}]}
{"id": "video_600__q0000", "video_id": "video_600", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_600_t5s.mp4"], "extra_info": {"question": "What is the relationship between the man in the blue top and who claims to have a strong personality?"}, "reward_model_ground_truth": {"multi_choice": ["A.first love", "B.crony", "C.elder sister and younger brother", "D.childhood sweetheart"], "correct_option": "D"}, "answer_text": "childhood sweetheart", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:05, the lady said 'have a strong personality'.", "inference": "Get the lady they talked about."}, {"modality": "vision", "evidence": "At 2:05, the lady was called rich lady.", "inference": "Her character is rich lady."}, {"modality": "vision", "evidence": "At 2:14, the man in the blue is lawyer.", "inference": "His character is lawyer."}, {"modality": "audio", "evidence": "At 4:53, someone said the relationship of the rich lady and the lawyer is childhood sweetheart.", "inference": "So they are childhood sweetheart."}]}
{"id": "video_600__q0001", "video_id": "video_600", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_600_t5s.mp4"], "extra_info": {"question": "What is the identity of the person claiming destiny is in his own hands second?"}, "reward_model_ground_truth": {"multi_choice": ["A.designer", "B.CEO", "C.steward", "D.childe"], "correct_option": "A"}, "answer_text": "designer", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:52, locate the second phrase about 'My destiny is in my own hands'.", "inference": "Get the character they talked about."}, {"modality": "vision", "evidence": "Observe the information on the character's screen to infer her identity.", "inference": "She is designer."}]}
{"id": "video_601__q0000", "video_id": "video_601", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_601_t5s.mp4"], "extra_info": {"question": "What is the meaning of the narration for the third laugh?"}, "reward_model_ground_truth": {"multi_choice": ["A. Ross kept getting interrupted", "B. Ross simplified the blessing into one word", "C. The waiter needed to return to his seat to rest first", "D. The waiter questioned Joey's absurd assumption"], "correct_option": "B"}, "answer_text": "Ross simplified the blessing into one word", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:31, Ross is introducing Monica with long blessing.", "inference": "Introduce the background of the scene."}, {"modality": "aduio", "evidence": "At 0:33, during the narration, laughing appears the 3rd time.", "inference": "This is the third time of laughing."}, {"modality": "vision", "evidence": "At 0:33, Ross simply lifted the cup upward rapidly", "inference": "He wants to hurry up with the blessing, preventing another interruption."}]}
{"id": "video_601__q0001", "video_id": "video_601", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_601_t5s.mp4"], "extra_info": {"question": "Where is Monica sitting?"}, "reward_model_ground_truth": {"multi_choice": ["A.On the left side of Rachel", "B.On the left side of Joey", "C.On the left side of the blonde lady in black", "D.On the left side of the man wearing a tie"], "correct_option": "D"}, "answer_text": "On the left side of the man wearing a tie", "reasoning_steps": [{"modality": "audio", "evidence": "At the very beginning, the man wearing a tie introduces Monica.", "inference": "Remind people to notice the question."}, {"modality": "audio", "evidence": "During the man's introduction, the woman keeps adding information about Monica.", "inference": "This woman is Monica."}, {"modality": "vision", "evidence": "Monica sits on the left side of the man wearing a tie.", "inference": "This is the place where Monica is."}]}
{"id": "video_602__q0000", "video_id": "video_602", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_602_t5s.mp4"], "extra_info": {"question": "What mainly caused the most exclamation marks to appear on the screen?"}, "reward_model_ground_truth": {"multi_choice": ["A. Bai Lu's acting is very impressive.", "B. Li Chen flipped the table very far.", "C. Fan Chengcheng flipped the table very far.", "D. Zhang Zhenyuan's acting is very impressive."], "correct_option": "B"}, "answer_text": "Li Chen flipped the table very far.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:31, the man named Li Chen throws the table far away.", "inference": "Li Chen throws the table in a extremely far distance."}, {"modality": "audio", "evidence": "At 0:33, other players say 'Wow'.", "inference": "Other players are shocked."}, {"modality": "vision", "evidence": "There are many exclamation marks and players' face appearing on the screen.", "inference": "Many players are shocked by Li Chen's amazing actions of throwing things far away."}]}
{"id": "video_602__q0001", "video_id": "video_602", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_602_t5s.mp4"], "extra_info": {"question": "What identity was the character who was asked 'How much did this cost?'"}, "reward_model_ground_truth": {"multi_choice": ["A.CEO", "B.Housewife", "C.Captain", "D.Designer"], "correct_option": "C"}, "answer_text": "Captain", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:33, A man asked another man 'How many items did you buy? How much did this cost'.", "inference": "It is the moment mentioned in the question."}, {"modality": "audio", "evidence": "At 3:35, a man says 'Not much. Several hundred'.", "inference": "This man was answer the question 'how much did this cost'."}, {"modality": "vision", "evidence": "At 3:35, the tag behind the man answering price question writes ' Zhang Zhenyuan Captain'.", "inference": "The character who was asked 'How much did this cost' is a captain."}]}
{"id": "video_602__q0002", "video_id": "video_602", "qid": 2, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_602_t5s.mp4"], "extra_info": {"question": "Who plays the role of the son"}, "reward_model_ground_truth": {"multi_choice": ["A.Fan Chengcheng", "B.Meng Ziyi", "C.Zhang Zhenyuan", "D.Ao Ruipeng"], "correct_option": "B"}, "answer_text": "Meng Ziyi", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:08,  a woman announces the title of the script 'a father and a son'.", "inference": "The script Includes the role of a father and a son."}, {"modality": "audio", "evidence": "a man says' Meng Ziyi, I play this script with you'.", "inference": "A man wants to play 'A father and a son' with Meng Ziyi."}, {"modality": "vision", "evidence": "At 1:29, there's a poster writes that Meng Ziyi plays the son and Fan Chengcheng plays the father.", "inference": "MengZiyi plays the role of the son."}]}
{"id": "video_603__q0000", "video_id": "video_603", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_603_t5s.mp4"], "extra_info": {"question": "Why do teachers teach Sum-to-Product Identities and Product-to-Sum Identities?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because Gaokao tested this knowledge.", "B.Because the knowledge was included in the textbook.", "C.Because many students asked about this knowledge.", "D.Because this knowledge is often overlooked."], "correct_option": "A"}, "answer_text": "Because Gaokao tested this knowledge.", "reasoning_steps": [{"modality": "vision", "evidence": "The question presented in Gaokao is about Sum-to-Product Identities.", "inference": "Introduce the topic of this video."}, {"modality": "audio", "evidence": "At 0:06, the teacher says that in the future, Sum-to-Product Identities will become the compulsory course.", "inference": "There will be more questions about this knowledge in Gaokao."}]}
{"id": "video_603__q0001", "video_id": "video_603", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_603_t5s.mp4"], "extra_info": {"question": "Which part of the 'Magic Little Pig' icon is animated?"}, "reward_model_ground_truth": {"multi_choice": ["A.Ears.", "B.Tail.", "C.Eyes.", "D.Nose."], "correct_option": "D"}, "answer_text": "Nose.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:01, teacher introduces himself as 'Magic Little Pig'.", "inference": "Remind people to notice the beginning part."}, {"modality": "vision", "evidence": "At 0:01, the 'Magic Little Pig' icon appears on the screen.", "inference": "Mark the very point of the question."}, {"modality": "vision", "evidence": "The nose of 'Magic Little Pig' icon is moving.", "inference": "The nose of the 'Magic Little Pig' icon is animated."}]}
{"id": "video_604__q0000", "video_id": "video_604", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_604_t5s.mp4"], "extra_info": {"question": "Before the 'toe-touch side plank' is mentioned in the audio, how many dynamic plank variations has the instructor demonstrated in total?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1 variation.", "B. 4 variations.", "C. 3 variations.", "D. 2 variations."], "correct_option": "A"}, "answer_text": "1 variations.", "reasoning_steps": [{"modality": "audio", "evidence": "The narrator introducing exercise #2 as 'Tom Cruise's plank' at timestamp 03:04.", "inference": "The time of the plank is introduced in time 03:04."}, {"modality": "audio", "evidence": "The instructor first demonstrates the 'Classic Plank' at 00:37 .", "inference": "Classic Plank is the first kind of plank variations."}, {"modality": "audio", "evidence": "The narrator introduces exercise 'Push-up Plank'  at 01:52 .", "inference": "Push-up Plank is the second kind of plank variations."}, {"modality": "vision", "evidence": "'Tom Cruise's plank' requires movement.  at 03:34 .", "inference": "'Tom Cruise's plank' is a dynamic plank variation."}]}
{"id": "video_604__q0001", "video_id": "video_604", "qid": 1, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_604_t5s.mp4"], "extra_info": {"question": "For the first numbered exercise shown, why is the 'Push-up Plank' designed this way?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because it's a moderate-intensity aerobic exercise that gets your heart rate up.", "B. Because it's a modified, more accessible version of the pushup for beginners.", "C. Because it can help prevent injuries during endurance training.", "D. Because it primarily targets the glutes and latissimus dorsi."], "correct_option": "B"}, "answer_text": "Because it's a modified, more accessible version of the pushup for beginners.", "reasoning_steps": [{"modality": "vision", "evidence": "The woman's visual action strongly resembles a modified push-up.", "inference": "The 'Push-up Plank' is same with the woman's action ."}, {"modality": "audio", "evidence": "The narrator states, it's a lot like a push-up, it's just been modified to make it easier for beginners to get started  from 02:04 to 02:09", "inference": "It's designed for the beginners."}]}
{"id": "video_605__q0000", "video_id": "video_605", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_605_t5s.mp4"], "extra_info": {"question": "What triage category would a patient who is 'grabbing their chest in pain' be assigned to, and within what timeframe do they need to be seen?"}, "reward_model_ground_truth": {"multi_choice": ["A. Category 1, to be seen within 2 minutes.", "B. Category 5, to be seen when the condition is less time-pressured.", "C. Category determined within 3 minutes by the triage nurse.", "D. Category depends on whether they arrived by ambulance or family."], "correct_option": "A"}, "answer_text": "Category 1, to be seen within 2 minutes.", "reasoning_steps": [{"modality": "audio", "evidence": "At 02:29, Alison gives a specific example: ''if you're grabbing your chest, curling over in pain, we can see that immediately''. This establishes that this symptom is considered highly urgent.", "inference": "Alison explains the patient who is grabbing their chest."}, {"modality": "audio", "evidence": "At 02:56, Dr. William, explains the Australasian Triage Scale. He states, ''Category 1 will be patients that need to be seen immediately and that they have a life-threatening problem.''", "inference": "Category 1 will be patients that need to be seen immediately."}, {"modality": "vision", "evidence": "At 04:33, text on screen states, ''THE SICKEST PATIENTS IN OUR EMERGENCY DEPARTMENTS ARE SEEN WITHIN 2 MINUTES OF ARRIVING.''", "inference": "These sickest patients must be seen within 2 minutes."}]}
{"id": "video_606__q0000", "video_id": "video_606", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_606_t5s.mp4"], "extra_info": {"question": "As the two individuals are being arrested at the airport, what was found on them?"}, "reward_model_ground_truth": {"multi_choice": ["A. Illegal items hidden in their luggage.", "B. Some loose clothes used for camouflage.", "C. Packages of heroin tightly taped to their bodies.", "D. Some counterfeit passports and documents."], "correct_option": "C"}, "answer_text": "Packages of heroin tightly taped to their bodies.", "reasoning_steps": [{"modality": "vision", "evidence": "At timestamp 2:04, the video shows two individual being escorted by officials in an airport.", "inference": "They are the two individuals arrested at the airport in the question."}, {"modality": "audio", "evidence": "Immediately following this scene, at 2:09, the narrator explains they were hiding 'packages of heroin tightly taped to their bodies'.", "inference": "The two individuals were hiding packages of heroin tightly taped to their bodies."}, {"modality": "vision", "evidence": "At 2:09, photographs show packages of drugs being removed from their body.", "inference": "By combining the visual of the arrest with the narrator's specific description of the hidden drugs and the photographic evidence, it is clear that the reason for their arrest was the discovery of heroin taped to their bodies."}]}
{"id": "video_606__q0001", "video_id": "video_606", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_606_t5s.mp4"], "extra_info": {"question": "When Renae Lawrence's picture is first shown, what does the audio say is the context of her actions?"}, "reward_model_ground_truth": {"multi_choice": ["A. She was in Bali to transport drugs.", "B. She was in Bali to smuggle drugs.", "C. She was on a journey to Bali to smuggle drugs.", "D. She was in Indonesia to smuggle drugs."], "correct_option": "B"}, "answer_text": "She was in Bali to smuggle drugs.", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:03, the video introduced Renae Lawrence for the first time by name.", "inference": "Here we know who Renae Lawrence is."}, {"modality": "audio", "evidence": "At 0:22, the narrator states that nine young Australians had been recruited to smuggle illegal substances out of Bali.", "inference": "In the video, there are nine young people recruited to smuggle illegal substances out of Bali."}, {"modality": "vision", "evidence": "At 2:09, the video shows Lawrence was hiding heroin.", "inference": "By linking the introduction of Renae Lawrence to the broader criminal plot already described by the narrator, we can infer that she was one of the nine young Australians and her trip to Bali was for the purpose of smuggling drugs."}]}
{"id": "video_607__q0000", "video_id": "video_607", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_607_t5s.mp4"], "extra_info": {"question": "When the practicality section is mentioned, what is the initial action of the person in the video?"}, "reward_model_ground_truth": {"multi_choice": ["A.Mentioned the robot vacuum has a certain odor, then showed its front, which has two wheels and two round cleaning pads", "B.Mentioned it has a deodorizing effect, then showed the back of the robot, which has two wheels and three round cleaning pads", "C.Mentioned the robot vacuum has a deodorizing effect, then showed its back, which has two wheels and two round cleaning pads", "D.Mentioned that it has a certain odor, then showed the back of the robot vacuum, which has two wheels and two round cleaning pads"], "correct_option": "D"}, "answer_text": "Mentioned that it has a certain odor, then showed the back of the robot vacuum, which has two wheels and two round cleaning pads", "reasoning_steps": [{"modality": "vision", "evidence": "At 6:50, the introducing title in the center of the picture shows: Practicality in Chinese.", "inference": "Sets the point to check the initial action of the person."}, {"modality": "audio", "evidence": "At 7:12, the reviewer said, 'Well, it's impossible to be fragrant, it does have a certain odor.'", "inference": "Mentioned that it has a certainj odor."}, {"modality": "vision", "evidence": "At the same moment, the reviewer flips the robot vacuum over to show its back.", "inference": "The reviewer showed its back."}, {"modality": "vision", "evidence": "Verify the visual details of the robot's back. The video shows it has two main drive wheels and two round spinning cleaning pads.", "inference": "In the end,  the video showed its two wheels and two round cleaning pads。"}]}
{"id": "video_607__q0001", "video_id": "video_607", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_607_t5s.mp4"], "extra_info": {"question": "Which of the following statements does not represent a suggestion or critique the reviewer made regarding the DJI robot vacuum?"}, "reward_model_ground_truth": {"multi_choice": ["A. Incorporating a detachable sealing mechanism", "B. Delivering a user experience inferior to competing brands", "C. Enhancing the system's intelligent voice interaction functions", "D. Positioning it as a groundbreaking and transformative product"], "correct_option": "B"}, "answer_text": " Delivering a user experience inferior to competing brands", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:40, the video starts to show the special cameras and radars on the robot.", "inference": "The reviewer might think these cameras and radars are something special about this vacuum."}, {"modality": "audio", "evidence": "At 3:01, the reviewer's voice mentioned, 'So from the perspective of experience...'", "inference": "Sets the point to check if the reviewer thinks the vacuum is inferior or not."}, {"modality": "audio", "evidence": "At 3:05, the reviewer explicitly said that he thinks this is what makes DJI vacuum most special and most competitive comparing to other product.", "inference": "User's experience is superior instead of inferior to competing brands."}]}
{"id": "video_608__q0000", "video_id": "video_608", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_608_t5s.mp4"], "extra_info": {"question": "What are the auxiliary means to achieve the main purpose of the video?"}, "reward_model_ground_truth": {"multi_choice": ["A.To share a journey through London.", "B.To teach English vocabulary to learners.", "C.To visit locations such as the London Eye and other famous places.", "D.To point out several special words in the English for items."], "correct_option": "D"}, "answer_text": "To point out several special words in the English for items.", "reasoning_steps": [{"modality": "audio", "evidence": "At the beginning of the video, the main character explicitly states this goal, saying, 'So let's travel and learn some English.", "inference": "He is teching some main vocabularies."}, {"modality": "vision", "evidence": "Throughout the trip, the main character points out various objects and landmarks, and at the same time,the corresponding English words appear on screen.", "inference": "point out several special words in the English for items."}]}
{"id": "video_608__q0001", "video_id": "video_608", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_608_t5s.mp4"], "extra_info": {"question": "What is the role of Uncle Lee?"}, "reward_model_ground_truth": {"multi_choice": ["A.Uncle Lee is the man who goes the same way as the main character.", "B.Uncle Lee is the man seen travelling with the speaker throughout London.", "C.Uncle Lee is the man who sends the main character on a train to his hometown of York.", "D.Uncle Lee acts as a local guide for the London portion."], "correct_option": "D"}, "answer_text": "Uncle Lee acts as a local guide for the London portion.", "reasoning_steps": [{"modality": "audio", "evidence": "The main character calls the man who accompanies him 'Uncle Lee'.", "inference": "The man is called 'Uncle Lee'."}, {"modality": "vision", "evidence": "The main character arrives in lodon,and then he meets a man who accompanies him to have a trip around London.", "inference": "Uncle Lee acts as a locl guide for London."}]}
{"id": "video_609__q0000", "video_id": "video_609", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_609_t5s.mp4"], "extra_info": {"question": "What does the background music indicate when shopping?"}, "reward_model_ground_truth": {"multi_choice": ["A.She likes the thing in the shop.", "B.She is enjoyable to find a cute color.", "C.She enjoys making a day-in-the-life vlog", "D.She likes the music in the shopping mall."], "correct_option": "B"}, "answer_text": "She is enjoyable to find a cute color.", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:28, she says I'm going to see if I can find a cute color before enter the shopping mall.", "inference": "the background music is joyful."}, {"modality": "vision", "evidence": "At 4:28, the protagonist wears smile when shopping.", "inference": "She is enjoyable to find a cute color."}]}
{"id": "video_609__q0001", "video_id": "video_609", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_609_t5s.mp4"], "extra_info": {"question": "In which direction will Drew think the protagonist is moving?"}, "reward_model_ground_truth": {"multi_choice": ["A.Right.", "B.Left.", "C.Forward", "D.Backward."], "correct_option": "B"}, "answer_text": "Left.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:40, The woman says that she will meet a friend called Drew, who is working in starbucks.", "inference": "Get the point they talked about."}, {"modality": "vision", "evidence": "The woman found her friend, and then she will drive ahead.", "inference": "That is Drew's left."}]}
{"id": "video_61__q0000", "video_id": "video_61", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_61_t5s.mp4"], "extra_info": {"question": "What tone does the woman use when talking about the grapes"}, "reward_model_ground_truth": {"multi_choice": ["A. A tone of deep apology", "B. An angry tone", "C. A tone of disdain / a contemptuous tone", "D. An awkward tone"], "correct_option": "B"}, "answer_text": "An angry tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:01, A woman appeared on the screen", "inference": "Confirm whether she is the target woman. "}, {"modality": "audio", "evidence": "At 00:04, the woman asked, 'Why would you eat the grapes?'", "inference": "This indicates that the woman was very angry with the man for eating the grapes."}]}
{"id": "video_61__q0001", "video_id": "video_61", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_61_t5s.mp4"], "extra_info": {"question": "What result happened to the boy's tone of voice when he sees the man trying to avoid the piano performance?"}, "reward_model_ground_truth": {"multi_choice": ["A. The boy Slightly sarcastic to it", "B. The boy pretends to be sad", "C. The boy truly sad", "D. The boy disappointed"], "correct_option": "A"}, "answer_text": "The boy Slightly sarcastic to it", "reasoning_steps": [{"modality": "vision", "evidence": "At 07:35,A man, out of concern, wanted to avoid playing the piano for a while first. ", "inference": "Key information captured."}, {"modality": "audio", "evidence": "At 07:35, The boy looked at the man who was trying to run away and said, 'Convenient' ", "inference": "This indicates that the boy Slightly sarcastic tothe man for his evasive behavior."}]}
{"id": "video_610__q0000", "video_id": "video_610", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_610_t5s.mp4"], "extra_info": {"question": "What was ordered at the store of the non-English person who spoke in the video? "}, "reward_model_ground_truth": {"multi_choice": ["A. Pork stock.", "B. Slender noodles.", "C. Dumplings.", "D. Several eggs."], "correct_option": "D"}, "answer_text": "Several eggs", "reasoning_steps": [{"modality": "audio", "evidence": "At 7:59, a woman said in Chinese, 'This is the oldest type of noodle in our area.'", "inference": "Sets the point to check what was ordered."}, {"modality": "vision", "evidence": "At 8:05, there are three eggs appeared in the vloggers' table.", "inference": "Several eggs were ordered."}]}
{"id": "video_610__q0001", "video_id": "video_610", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_610_t5s.mp4"], "extra_info": {"question": "When someone claim China does the best stocks in the world, which seat does the man wearing a baseball cap and glasses sit in?"}, "reward_model_ground_truth": {"multi_choice": ["A.The first", "B.The second", "C.The third", "D.The fourth"], "correct_option": "B"}, "answer_text": "The second", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:57, a male voice said that he honestly thinks China does the best stocks in the world.", "inference": "Sets the point to check the visuals."}, {"modality": "vision", "evidence": "At the same moment, the man wearing a baseball cap and glassess sitting in the second seat.", "inference": "Confirm that the man sits in the second seat."}]}
{"id": "video_611__q0000", "video_id": "video_611", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_611_t5s.mp4"], "extra_info": {"question": "Why was a character briefly surrounded by flames?"}, "reward_model_ground_truth": {"multi_choice": ["A. He was tricked into an unsafe place.", "B. His teammate refused boys helping boys.", "C. His teammate mocked him as like Ma Boqian.", "D. He fell into the water in two rounds."], "correct_option": "A"}, "answer_text": "He was tricked into an unsafe place.", "reasoning_steps": [{"modality": "vision", "evidence": "At 10:43, Zhang Zhenyuan is surrounded by flames.", "inference": "Mark the point about the question that needs attention."}, {"modality": "audio", "evidence": "At 7:49, Song Yuqi asks Zhang Zhenyuan to change their seats and Zhang Zhenyuan agrees.", "inference": "Song Yuqi is 2, Zhang Zhenyuan is 1."}, {"modality": "vision", "evidence": "At 8:36, 1 is ejected.", "inference": "If not change, Zhang Zhenyuan is 2 and is safe as well, so his is angry."}]}
{"id": "video_611__q0001", "video_id": "video_611", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_611_t5s.mp4"], "extra_info": {"question": "With whom had Song Yuqi previously made an agreement that led her to complain about being ejected?"}, "reward_model_ground_truth": {"multi_choice": ["A. Zhang Zhenyuan.", "B. Fan Chengcheng.", "C. Song Yuqi.", "D. Meng Zhiyi."], "correct_option": "A"}, "answer_text": "Zhang Zhenyuan", "reasoning_steps": [{"modality": "vision", "evidence": "At 6:46, Song Yuqi is labelled as team captain.", "inference": "Mark the point about the question that needs attention."}, {"modality": "audio", "evidence": "At 7:52, Song Yuqi and Zhang Zhenyuan reached a consensus", "inference": "Zhang Zhenyuan agreed not to let Song Yuqi be catapulted out in the next round."}, {"modality": "audio", "evidence": "At 11:15,As the team leader himself, Zhang Zhenyuan complained that he had been catapulted out too—and Song Yuqi was also catapulted out.", "inference": "Song Yuqi complained that she had been catapulted out with Zhang Zhenyuan"}]}
{"id": "video_611__q0002", "video_id": "video_611", "qid": 2, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_611_t5s.mp4"], "extra_info": {"question": "When Fan Chengcheng was the team captain, which seat was he always recommending?"}, "reward_model_ground_truth": {"multi_choice": ["A.1", "B.2", "C.3", "D.4"], "correct_option": "D"}, "answer_text": "4", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:29, Fan Chengcheng is labelled as team captain.", "inference": "Mark the point about the question that needs attention."}, {"modality": "audio", "evidence": "At 0:47, Fan Chengcheng recommends 4 and 2 to the first team member.", "inference": "4 and 2 are recommneded for the first time."}, {"modality": "audio", "evidence": "At 1:39, Fan Chengcheng recommends 4 to the second team member.", "inference": "4 is recommended for the second time."}, {"modality": "audio", "evidence": "At 2:11, Fan Chengcheng recommends 4 to the third team member.", "inference": "4 is recommeded for the third time."}]}
{"id": "video_611__q0003", "video_id": "video_611", "qid": 3, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_611_t5s.mp4"], "extra_info": {"question": "Who would fall into the water if Song Yuqi's seat change request as the team leader was not accepted?"}, "reward_model_ground_truth": {"multi_choice": ["A. Zhang Zhenyuan and Fan Chengcheng.", "B. Meng Ziyi and Zhang Zhenyuan.", "C. Song Yuqi and Meng Ziyi.", "D. Fan Chengcheng and Song Yuqi."], "correct_option": "D"}, "answer_text": " Fan Chengcheng and Song Yuqi.", "reasoning_steps": [{"modality": "vision", "evidence": "At 6:46, Song Yuqi is labelled as team captain.", "inference": "Mark the point about the question that needs attention."}, {"modality": "vision", "evidence": "At 7:05, Meng Ziyi chooses 3.", "inference": "Mark Meng Ziyi's seat."}, {"modality": "vision", "evidence": "At 7:22, Fan Chengcheng chooses 4.", "inference": "Mark Fan Chengcheng's seat."}, {"modality": "vision", "evidence": "At 7:35, Zhang  Zhenyuan chooses 2.", "inference": "Mark Zhang Zhenyuan's seat."}, {"modality": "vision", "evidence": "At 7:35, Song Yuqi chooses 1.", "inference": "Mark Song Yuqi's seat."}, {"modality": "audio", "evidence": "At 7:36, Fan Chengcheng asks Meng Ziyi to change their seats, but Meng Ziyi refuses.", "inference": "Fan Chengcheng is 4, Meng Ziyi is 3."}, {"modality": "audio", "evidence": "At 7:49, Song Yuqi asks Zhang Zhenyuan to change their seats and Zhang Zhenyuan agrees.", "inference": "Song Yuqi is 2, Zhang Zhenyuan is 1."}, {"modality": "vision", "evidence": "At 8:36, 1 and 4 are ejected.", "inference": "If not change, 1 is Song Yuqi and 4 is Fan Chengcheng."}]}
{"id": "video_611__q0004", "video_id": "video_611", "qid": 4, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_611_t5s.mp4"], "extra_info": {"question": "When Meng Zhiyi was the team leader, for the first person who sat in the elimination seat, when were they first advised to take that seat?"}, "reward_model_ground_truth": {"multi_choice": ["A.Meng Zhiyi ", "B.Song Yuqi ", "C.Zhang Zhenyuan ", "D.Fan Chengcheng "], "correct_option": "D"}, "answer_text": "Fan Chengcheng", "reasoning_steps": [{"modality": "vision", "evidence": "At 4:10, Meng Ziyi is labelled as team captain.", "inference": "Mark the point about the question that needs attention."}, {"modality": "vision", "evidence": "At 4:13, Meng Ziyi chooses 1 and 2 that will be ejected.", "inference": "Note the members who select these two seats."}, {"modality": "vision", "evidence": "At 5:38, Fan Chengcheng chooses 2.", "inference": "Fan Chengcheng will be ejected and he was the first person who sat in the elimination seat "}, {"modality": "audio", "evidence": "At 4:35, Meng Ziyi advised Fanc Chengcheng to take the 1st seat for the  1st advice.", "inference": "Note the first advice "}]}
{"id": "video_612__q0000", "video_id": "video_612", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_612_t5s.mp4"], "extra_info": {"question": "Which material mentioned in the video is manufactured using the process shown starting at 11"}, "reward_model_ground_truth": {"multi_choice": ["A. 440C stainless steel", "B. Silicon nitride", "C. High-speed tool steel alloy/tungsten carbide", "D. Diamond-like carbon"], "correct_option": "C"}, "answer_text": "High-speed tool steel alloy/tungsten carbide", "reasoning_steps": [{"modality": "vision", "evidence": "At 11:20 the footage shows molten metal being atomized into a fine powder, which is then pressed into shapes. This process is visually distinct.", "inference": "Molten metal is atomized into a fine powder."}, {"modality": "vision", "evidence": "At 12:25, the interior of a rapidly spinning jet engine turbine, a clear example of a high-temperature, high-speed environment.", "inference": "an example suitable for high-temperature, high-speed applications."}, {"modality": "audio", "evidence": "At 11:20,  the narrator identifies the manufacturing process shown as the 'powder metallurgy process' used to make 'carbides'.", "inference": "the pressing process is called powder metallurgy, and the resulting material is carbides."}, {"modality": "audio", "evidence": "Later, when discussing high-performance applications, the narrator mentions that materials like carbides and ceramics are necessary for extreme temperatures where steel would fail.", "inference": "carbides are the specific material group, made via the previously shown process, that is suited for high-temperature and high-speed conditions."}]}
{"id": "video_613__q0000", "video_id": "video_613", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_613_t5s.mp4"], "extra_info": {"question": "What explains the 'French Paradox'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Red wine can offset the negative cardiovascular effects of a high-fat diet.", "B. Red wine can counteract harmful cardiac effects of fatty foods.", "C. Red wine can offset the negative heart effects of a high-fat diet.", "D. Red wine can neutralize adverse heart impacts of high-fat consumption."], "correct_option": "B"}, "answer_text": " Red wine can counteract harmful cardiac effects of fatty foods.", "reasoning_steps": [{"modality": "vision", "evidence": "The television shows that the text 'GLASS OF RED WINE EQUIVALENT TO HOUR IN GYM TIME' in time 60s .", "inference": "Red wine could be good for human body."}, {"modality": "audio", "evidence": "The narrator of the '60 Minutes' segment poses the riddle: 'Why don't French people have more heart attacks if they eat so much cheese?'.", "inference": "French people eats a lot of cheese."}, {"modality": "vision", "evidence": "The graph shows that the lowest spot of morality rate of 2 drinks .", "inference": "People drinks 2 wines have less possibility of having cancer and disease."}, {"modality": "vision", "evidence": "The TV shows that the alcohol reduces the risk of heart disease .", "inference": "Red wine could protect the heart."}]}
{"id": "video_613__q0001", "video_id": "video_613", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_613_t5s.mp4"], "extra_info": {"question": "How did alcohol guidelines in Australia and Canada change chronologically and in severity?"}, "reward_model_ground_truth": {"multi_choice": ["A. Australia (2020): 10 drinks/week. Canada (2023): more restrictive 2 drinks/week.", "B. Canada (2020): 10 drinks/week. Australia (2023): stricter 2 drinks/week.", "C. Australia (2020): 10 drinks/week. Canada (2023): stricter 2 drinks/week.", "D. Australia (2020): 2 drinks/week. Canada (2023): less restrictive 10 drinks/week."], "correct_option": "C"}, "answer_text": "Australia (2020): 10 drinks/week. Canada (2023): stricter 2 drinks/week.", "reasoning_steps": [{"modality": "vision", "evidence": "The video displays a TV screen with the text 'Australia 2020'.", "inference": "The news happed in Australia in 2020."}, {"modality": "audio", "evidence": "A voiceover says adults should have a maximum of 10 standard drinks a week.", "inference": "Australia's guideline change to 10 drinks a week."}, {"modality": "vision", "evidence": "The video displays a TV screen with the text 'Canada 2023'.", "inference": "The news happed in Canada in 2023."}, {"modality": "audio", "evidence": "A voiceover says no more than two drinks a week.", "inference": "Canada's guideline change to 2 drinks a week."}]}
{"id": "video_614__q0000", "video_id": "video_614", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_614_t5s.mp4"], "extra_info": {"question": "Where were they preparing to set off for when they were urged by the lion?"}, "reward_model_ground_truth": {"multi_choice": ["A. The Empire State Building.", "B. The Statue of Liberty.", "C. The Time Square.", "D. The Hollywood"], "correct_option": "C"}, "answer_text": "The Time Square.", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:03, there was a horn sound from the car behind to urge them.", "inference": "Set the point to check if the car is driven by a lion."}, {"modality": "vision", "evidence": "At 3:05, the lion was driving the car behind them.", "inference": "The lion was urging them at this point."}, {"modality": "audio", "evidence": "At 3:08, Peppa Pig asked, 'Where are we going next? Miss Rabbit.'", "inference": "Pay attention to what the Miss Rabbit answers about their destination."}]}
{"id": "video_614__q0001", "video_id": "video_614", "qid": 1, "question_type": "spatial understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_614_t5s.mp4"], "extra_info": {"question": "What was the standing position of Miss Rabbit, who had sung a song, in the filming studio?"}, "reward_model_ground_truth": {"multi_choice": ["A. In the first position.", "B. In the second position.", "C. In the third position.", "D. In the fourth position."], "correct_option": "C"}, "answer_text": "In the third position.", "reasoning_steps": [{"modality": "audio", "evidence": "At 5:35, a voice starts to sing.", "inference": "Set the point to check who is the one singing at the moment."}, {"modality": "vision", "evidence": "At the momemt, it is the Miss Rabbit with brown hat and clothes in the video.", "inference": "It's the Miss Rabbit with brown hat and clothes who is singing."}, {"modality": "vision", "evidence": "At 17:01, 4 Miss Rabbits were standing. Among them, the 3rd one is the Miss Rabbit with brown hat and clothes who has sung a song before.", "inference": "So Miss Rabbit, who had sung a song, standed in the 3rd position."}]}
{"id": "video_614__q0002", "video_id": "video_614", "qid": 2, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_614_t5s.mp4"], "extra_info": {"question": "What was Daddy Pig's breakfast on the second day in America?"}, "reward_model_ground_truth": {"multi_choice": ["A. Flipped carrots.", "B. Unflipped carrots.", "C. Flipped fried eggs.", "D. Unflipped fried eggs."], "correct_option": "D"}, "answer_text": "Unflipped fried eggs.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:01, the title shows America in the center of the picture.", "inference": "This episode of Peppa Pig could be about the story happening in America."}, {"modality": "audio", "evidence": "At 0:44, the Mommy Pig said, 'It means we're going to America.'", "inference": "Confirm that what's happening next is in America."}, {"modality": "vision", "evidence": "At 1:19, there is a American national flag on the left of the screen and the whole family arived.", "inference": "They were already in America at this point."}, {"modality": "audio", "evidence": "At 4:28, the narrator said that this family needs a good night sleep for their journey ahead.", "inference": "Their 1st day in America ends."}, {"modality": "audio", "evidence": "At 5:26, Daddy Pig said, 'I suppose we could stop for breakfast.'", "inference": "They're about to have the breakfast in America."}]}
{"id": "video_615__q0000", "video_id": "video_615", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_615_t5s.mp4"], "extra_info": {"question": "Why does one character's face turn red and white alternately?"}, "reward_model_ground_truth": {"multi_choice": ["A. The party was too lively.", "B. She had just run for a long time.", "C. Peppa's family party was too anticipated.", "D. Peppa's baby carrots excuse was too funny."], "correct_option": "B"}, "answer_text": "She had just run for a long time.", "reasoning_steps": [{"modality": "audio", "evidence": "One forget to put the result in the ballon.", "inference": "She had to send the mail soon by running for a long time"}, {"modality": "vision", "evidence": "One's face turn red and white after she stopped running.", "inference": "Running for a long time cased her face turnning red and white."}]}
{"id": "video_615__q0001", "video_id": "video_615", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_615_t5s.mp4"], "extra_info": {"question": "When what George is learning is mentioned, how many animals are there in the frame?"}, "reward_model_ground_truth": {"multi_choice": ["A.3", "B.4", "C.5", "D.6"], "correct_option": "B"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:31,Peppa said that George was learning French.", "inference": "Find the corresponding time point."}, {"modality": "vision", "evidence": " At the corresponding time point, there were one white bear, one antelope, and two pink pigs", "inference": "A total of four animals."}]}
{"id": "video_616__q0000", "video_id": "video_616", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_616_t5s.mp4"], "extra_info": {"question": "What made George say 'Yuck' for the second time? "}, "reward_model_ground_truth": {"multi_choice": ["A. The cucumber handed over by Grandma.", "B. The cucumber handed over by Grandpa.", "C. The tomato handed over by Grandma.", "D. The tomato handed over by Mom."], "correct_option": "D"}, "answer_text": "The tomato handed over by Mom.", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:05, George said the 'Yuck' for the second time.", "inference": "Pay attention to what was happening at that time to find the reson for George to say 'Yuck'."}, {"modality": "vision", "evidence": "Mom gave George a tomato to eat, then George seemed he didn't like eating tomato.", "inference": "George said 'Yuck' because he didn't like the tomato handed over by Mom."}]}
{"id": "video_616__q0001", "video_id": "video_616", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_616_t5s.mp4"], "extra_info": {"question": "What fruits and vegetables were on the shelf to the left of the Onions when Peppa said 'here they are'?"}, "reward_model_ground_truth": {"multi_choice": ["A. lettuce", "B. radish", "C. eggplant", "D. pumpkin"], "correct_option": "C"}, "answer_text": "eggplant", "reasoning_steps": [{"modality": "audio", "evidence": "At 7:04, Peppa says the sentence 'here they are'.", "inference": "Pay attention to the fruit on the left of the Onions."}, {"modality": "vision", "evidence": "Look at the shelf, we can see that the vegetables on the left of the Onions were eggplant.", "inference": "Eggplants were on the shelf to the left of the Onions."}]}
{"id": "video_616__q0002", "video_id": "video_616", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_616_t5s.mp4"], "extra_info": {"question": "What happened just after Dad crossed out the second item on the list?"}, "reward_model_ground_truth": {"multi_choice": ["A. Peppa mentioned the third item on the list.", "B. George mentioned the third item on the list.", "C. Dad pointed out that the third item was onions.", "D. Dad pointed out that the third item was fruits."], "correct_option": "C"}, "answer_text": "Dad pointed out that the third item was onions.", "reasoning_steps": [{"modality": "vision", "evidence": "At 6:31, Dad crossed out the second item on the list.", "inference": "Pay attention what happened next."}, {"modality": "audio", "evidence": "Dad pointed out that the third item they needed to get was onions.", "inference": "After Dad crossed out the second item on the list, Dad pointed out that the third item was onions."}]}
{"id": "video_617__q0000", "video_id": "video_617", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_617_t5s.mp4"], "extra_info": {"question": "Who said 'a bowl of Gou Jian's Sword'?"}, "reward_model_ground_truth": {"multi_choice": ["A.Fan Chengcheng", "B.Meng Ziyi", "C.Zhang Zhenyuan", "D.Sha Yi"], "correct_option": "B"}, "answer_text": "Meng Ziyi", "reasoning_steps": [{"modality": "audio", "evidence": "At  7:49, a girl said the phrase 'a bowl of Gou Jian's Sword'.", "inference": "Confirm the voice that said the phrase."}, {"modality": "vision", "evidence": "The person who is speaking at that moment is Meng Ziyi.", "inference": "Meng Ziyi said the phrase."}]}
{"id": "video_617__q0001", "video_id": "video_617", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_617_t5s.mp4"], "extra_info": {"question": "Who were the three people in the first round officially?"}, "reward_model_ground_truth": {"multi_choice": ["A.Meng Ziyi, Sha Yi, Li Chen", "B.Zhang Zhenyuan, Sha Yi, Zheng Kai ", "C.Meng Ziyi, Sha Yi, Zhang Zhenyuan", "D.Zhang Zhenyuan, Sha Yi, Li Chen"], "correct_option": "D"}, "answer_text": "Zhang Zhenyuan, Sha Yi, Li Chen", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:31, the announcement says that it's time to kick off the first round.", "inference": "Set the point the check the three participants."}, {"modality": "vision", "evidence": "At 1:33, Zhang Zhenyuan raised hand to sign up for the first round.", "inference": "One of the members is Zhang Zhenyuan."}, {"modality": "audio", "evidence": "At 1:34, Sha Yi said, 'I'm in.'", "inference": "Another member is Sha Yi."}, {"modality": "vision", "evidence": "Meng Ziyi left her seat to join Zhang Zhenyuan and Sha Yi to form the first team.", "inference": "For now, the third person is Meng Ziyi。"}, {"modality": "vision", "evidence": "At 2:39, Meng Ziyi left the table.", "inference": "The third person of the fitst round is not her."}, {"modality": "vision", "evidence": "At 2:52, Li Chen joined the table to play the first round.", "inference": "Confirmed that the three people are Zhang Zhenyuan, Sha Yi, Li Chen."}]}
{"id": "video_618__q0000", "video_id": "video_618", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_618_t5s.mp4"], "extra_info": {"question": "How many distinct food items did the man acquire and/or consume before he said 'that's pretty good'."}, "reward_model_ground_truth": {"multi_choice": ["A. Four", "B. Five", "C. Six", "D. Seven"], "correct_option": "A"}, "answer_text": "Four", "reasoning_steps": [{"modality": "audio", "evidence": "at 0:04, the man bought baked potatoes.", "inference": "the first one."}, {"modality": "vision", "evidence": "at 0:05, he bought corn soup.", "inference": "the second one."}, {"modality": "vision", "evidence": "then, he bought frozen bananas.", "inference": "the third one."}, {"modality": "vision", "evidence": "after that, he bought fried meat.", "inference": "the fourth one."}, {"modality": "vision", "evidence": "after that, he bought fried meat.", "inference": "the fourth one."}, {"modality": "audio", "evidence": "he said 'that's pretty good'.", "inference": "Total four kind of food."}]}
{"id": "video_618__q0001", "video_id": "video_618", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_618_t5s.mp4"], "extra_info": {"question": "What is the spatial relationship and movement of the person in the white jacket relative to the camera as he discusses the frozen bananas?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is standing still, directly facing the camera and talking.", "B. He is walking forward, directly facing the camera and talking.", "C. He is walking sideways, looking at the camera intermittently while talking.", "D. He is walking backward, talking over his shoulder while facing away from the camera."], "correct_option": "B"}, "answer_text": "He is walking forward, directly facing the camera and talking.", "reasoning_steps": [{"modality": "vision", "evidence": "At the beginning of  the video, the person in the white jacket is quite far away from the camera, we can see his whole jacket.", "inference": "We can judge the original distance from camera to the person at the beginning."}, {"modality": "vision", "evidence": "When the person is biting the frozen banana, we can only see part of his jacket.", "inference": "This indicates he is walking forward,directing facing the camera."}, {"modality": "audio", "evidence": "He says 'I'm sorry about that',and so on.", "inference": "While he is walking forward, directly facing the camera, he is also talking."}]}
{"id": "video_619__q0000", "video_id": "video_619", "qid": 0, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_619_t5s.mp4"], "extra_info": {"question": "What is the primary relationship between the intense visual portrayal of the wildfires and the accompanying audio narration?"}, "reward_model_ground_truth": {"multi_choice": ["A. The narration highlights the unknown origins of the extensive fires shown, discussing various unconfirmed theories about their cause.", "B. The narration provides a detailed explanation of the fire suppression efforts, complementing the visual evidence of damage being contained.", "C. The narration speculates on the economic impact of the fires, while the visuals focus solely on the immediate destruction of properties.", "D. The narration explains the historical frequency of such wildfires in the region, offering context to the visually devastating scale."], "correct_option": "A"}, "answer_text": "The narration highlights the unknown origins of the extensive fires shown, discussing various unconfirmed theories about their cause.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:07, the narrator said 'the more theories and questions are gonna arise...'.", "inference": "It indicates that various unconfirmed theories about the wildfire cause has been said in the audio"}, {"modality": "vision", "evidence": "The video shows a terrifying scene of countless wildfires burning.", "inference": "Fear has led to all sorts of bizarre speculations about the fire and its origin."}]}
{"id": "video_62__q0000", "video_id": "video_62", "qid": 0, "question_type": "spatial understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_62_t5s.mp4"], "extra_info": {"question": "After simulating the lunar landing mission, what is the content of letters behind them?"}, "reward_model_ground_truth": {"multi_choice": ["A. United States", "B. Omega and United Statese", "C. Omega and Apollo", "D. Apollo and United States"], "correct_option": "A"}, "answer_text": "United States", "reasoning_steps": [{"modality": "audio", "evidence": "the landing simulation ends at 07:08.", "inference": "locate this point."}, {"modality": "vision", "evidence": "when they take a photo at 07:21, there are some letters.", "inference": "they are 'United States'."}]}
{"id": "video_62__q0001", "video_id": "video_62", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_62_t5s.mp4"], "extra_info": {"question": "170 days before launch, the woman recommended several brand partnerships to the man and showed brand posters. In the subsequent video, which of these proposed brands were not mentioned?"}, "reward_model_ground_truth": {"multi_choice": ["A. Tang and Omega", "B. Omega and Crest", "C. Fruit of the Loom and Crest", "D. Tang and Fruit of the Loom"], "correct_option": "C"}, "answer_text": "Fruit of the Loom and Crest", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:31, the man said ' what's else here?'.", "inference": "locate this point."}, {"modality": "vision", "evidence": "the woman recommended Omega, Crest toothpaste , Tang and Fruit of the Loom.", "inference": "get these brands."}, {"modality": "audio", "evidence": "At 5:24, 'Tang' is mentioned.", "inference": "get the 'Tang'."}, {"modality": "audio", "evidence": "At 5:44, 'Omega' is mentioned.", "inference": "get the 'Tang','Omega', no Fruit of the Loom and Crest."}]}
{"id": "video_620__q0000", "video_id": "video_620", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_620_t5s.mp4"], "extra_info": {"question": "Which musical genre and instrument are prominently featured by the character with the red guitar?"}, "reward_model_ground_truth": {"multi_choice": ["A. Funk Rock, Electric Guitar", "B. Acoustic Pop, Acoustic Guitar", "C. Heavy Metal, Electric Guitar", "D. Classical Rock, Piano"], "correct_option": "A"}, "answer_text": "Funk Rock, Electric Guitar", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:05, a female character holding an electric guitar appears.", "inference": "So the instrument mentioned in the question is the electric guitar."}, {"modality": "audio", "evidence": "At 0:08, a funk-rock style background music starts playing.", "inference": "So the musical genre mentioned in the question is Fuck Rock."}]}
{"id": "video_620__q0001", "video_id": "video_620", "qid": 1, "question_type": "relationship reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_620_t5s.mp4"], "extra_info": {"question": "What best describes the interaction between the two characters?"}, "reward_model_ground_truth": {"multi_choice": ["A. Character A is reluctantly asking to be let go, while Character B insists on engaging him with music.", "B. Character B is playfully inviting Character A to join a band, but Character A is too shy.", "C. Character A is confessing his foolishness to Character B, who responds with a song of encouragement.", "D. The two characters are performing a lighthearted musical act, with Character A's initial shyness transitioning to enjoyment."], "correct_option": "A"}, "answer_text": "Character A is reluctantly asking to be let go, while Character B insists on engaging him with music.", "reasoning_steps": [{"modality": "vision", "evidence": "At the very beginning of the video, Character A appears.", "inference": "Identify him."}, {"modality": "audio", "evidence": "At the very beginning of the video, the background music is signing '...Please let me go...'", "inference": "This indicates that Character A is reluctantly asking to be let go."}, {"modality": "vision", "evidence": "At 0:05, the Character B holding an electric guitar appears.", "inference": "Identify her."}, {"modality": "audio", "evidence": "At 0:05, background music is singing 'Come and get your love'.", "inference": "This indicates that Character B is trying to let Character A engage with music."}, {"modality": "audio", "evidence": "At 0:12, background music is singing 'Come and get your love' again.", "inference": "This indicates that Character B is insisting to let Character A engage with music."}]}
{"id": "video_621__q0000", "video_id": "video_621", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_621_t5s.mp4"], "extra_info": {"question": "What was the immediate consequence of Dark Willow's successful denial of the Radiant Lotus Pool from Spirit Breaker?"}, "reward_model_ground_truth": {"multi_choice": ["A. The Dire team immediately initiated a counter-push on the Radiant mid lane.", "B. Spirit Breaker managed to secure the Lotus Pool despite the interruption.", "C. The Spirit team, unable to secure the Lotus Pool, immediately redirected their efforts to destroy the Dire top lane tower.", "D. Dark Willow proceeded to secure a kill on Spirit Breaker due to her successful denial."], "correct_option": "C"}, "answer_text": "The Spirit team, unable to secure the Lotus Pool, immediately redirected their efforts to destroy the Dire top lane tower.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:04, the purple hero Dark willow shows up.", "inference": "To observe the hero's next move."}, {"modality": "vision", "evidence": "At 0:21, the Spirit team cannot get the reward from the Lotus Pool.", "inference": "The Spirit team is unable to secure the Lotus Pool"}, {"modality": "audio", "evidence": "At 0:22, a commentator says: No worries. Let's destroy the top tower.", "inference": "Spirit Team has decided to destory Dire team's top tower."}]}
{"id": "video_621__q0001", "video_id": "video_621", "qid": 1, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_621_t5s.mp4"], "extra_info": {"question": "After the host started counting numbers, what is the second events in this Dota 2 match clip?"}, "reward_model_ground_truth": {"multi_choice": ["A. FLCN destroyed Radian's top tower.", "B. FLCN destroyed Radian's mid tower.", "C. Team Spirit destroyed Dire's top tower.", "D. Team Spirit destroyed Dire's top tower."], "correct_option": "B"}, "answer_text": "FLCN destroyed Radian's mid tower.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:37, the host started counting numbers.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "At 0:41, the team named Tspirit has given the final blow and destroyed Dire's top tower.", "inference": "Team Spirit destroyed Dire's top tower."}, {"modality": "vision", "evidence": "At 0:45, the team named FLCN has given the final blow and destroyed Radian's mid tower.", "inference": "FLCN destroyed Radian's mid tower."}]}
{"id": "video_621__q0002", "video_id": "video_621", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_621_t5s.mp4"], "extra_info": {"question": "What is the precise sequence of major tower events depicted in this segment of the game?"}, "reward_model_ground_truth": {"multi_choice": ["A. The Radiant team engages Dire's top tower but does not destroy it, followed by the Dire team destroying Radiant's mid tower.", "B. The Radiant team successfully destroys Dire's top tower, which is then followed by the Dire team destroying Radiant's mid tower.", "C. The Dire team successfully destroys Radiant's mid tower, and then the Radiant team proceeds to destroy Dire's top tower.", "D. Both the Radiant and Dire teams simultaneously engage and destroy their respective opposite's top towers."], "correct_option": "C"}, "answer_text": "The Radiant team successfully destroys Dire's top tower, which is then followed by the Dire team destroying Radiant's mid tower.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:22, a man says: No worries. Let's destroy the top tower.", "inference": "To check which team is going to destroy which team's top tower."}, {"modality": "vision", "evidence": "At 0:24, the Radian team is gathering to destroy Dire team's top tower.", "inference": "To observe if they have destroyed it or not."}, {"modality": "audio", "evidence": "At 0:43, the game announcement says: Dire's top tower has been destroyed.", "inference": "Radian has successfully destroyed Dire's top tower."}, {"modality": "vision", "evidence": "At 0:45, the Dire team detroyed Radian's mid tower.", "inference": "The Dire team destroyed Radian's mid tower before the Radian destroyed their top tower."}]}
{"id": "video_621__q0003", "video_id": "video_621", "qid": 3, "question_type": "temporal understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_621_t5s.mp4"], "extra_info": {"question": "What is the main sequence of tower-related events?"}, "reward_model_ground_truth": {"multi_choice": ["A. Team Spirit destroys the top tower, followed by FLCN destroying the mid tower.", "B. FLCN destroys the top tower, followed by Team Spirit destroying the mid tower.", "C. Team Spirit destroys the mid tower, followed by FLCN destroying the top tower.", "D. Both teams focus on pushing bottom lane towers without significant tower destruction shown."], "correct_option": "A"}, "answer_text": " Team Spirit destroys the top tower, followed by FLCN destroying the mid tower.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:35, Radian team is attacking the Dire's top tower, which has little HP left.", "inference": "To observe closely which hero destroyed the tower."}, {"modality": "vision", "evidence": "At 0:37, the team named Tspirit has given the final blow and destroyed Dire's top tower.", "inference": "Team Spirit destroyed Dire's top tower."}, {"modality": "vision", "evidence": "At 0:42, Dire team is attacking the Radian's mid tower, which has little HP left.", "inference": "To observe closely which hero destroyed the tower."}, {"modality": "vision", "evidence": "At 0:43, the team named FLCN has given the final blow and destroyed Radian's mid tower.", "inference": "FLCN destroyed Radian's mid tower."}]}
{"id": "video_622__q0000", "video_id": "video_622", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_622_t5s.mp4"], "extra_info": {"question": "Before she said 'I also use some brown...', how many distinct colored materials are used to construct the pencil-shaped gift card holder?"}, "reward_model_ground_truth": {"multi_choice": ["A. 4", "B. 5", "C. 6", "D. 7"], "correct_option": "A"}, "answer_text": "4", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:15, we can hear she mentioned that.", "inference": "Identify the timestamp."}, {"modality": "vision", "evidence": "At 0:03, she used yellow materials.", "inference": "It's the first material."}, {"modality": "vision", "evidence": "At 0:04, she used brown materials.", "inference": "It's the second material."}, {"modality": "vision", "evidence": "At 0:06, she used pink materials.", "inference": "It's the third material."}, {"modality": "vision", "evidence": "At 0:13, she used grey materials.", "inference": "It's the fourth material."}]}
{"id": "video_624__q0000", "video_id": "video_624", "qid": 0, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_624_t5s.mp4"], "extra_info": {"question": "Given the intense, final moments of the game and the commentary, what would have been the most immediate consequence if Harrison Barnes' game-winning shot had failed to go in?"}, "reward_model_ground_truth": {"multi_choice": ["A. The Golden State Warriors would have secured a one-point victory as the clock expired.", "B. The Sacramento Kings would have faced a close defeat, ending their chances of winning.", "C. The game would have immediately proceeded into an overtime period to determine the winner.", "D. The shot would have been subject to instant replay review for a potential offensive foul."], "correct_option": "C"}, "answer_text": "The game would have immediately proceeded into an overtime period to determine the winner.", "reasoning_steps": [{"modality": "audio", "evidence": "The host mentioned Harrison Barnes at 0:22.", "inference": "Identify the correct game of Harrison Barnes."}, {"modality": "vision", "evidence": "The countdown has only two seconds left.", "inference": "The match is almost over."}, {"modality": "vision", "evidence": "The score turns into 114:111 at the end.", "inference": "The game would have immediately proceeded into an overtime period if Harrison Barnes failed."}]}
{"id": "video_625__q0000", "video_id": "video_625", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_625_t5s.mp4"], "extra_info": {"question": "What was the primary reason for the exuberant celebration by the ski jumper and his team immediately following his jump?"}, "reward_model_ground_truth": {"multi_choice": ["A. He set a new world record for the longest ski flight.", "B. He achieved the longest jump of the competition so far.", "C. He secured a first-place finish in the event.", "D. He exceeded the required distance for qualification."], "correct_option": "A"}, "answer_text": "He set a new world record for the longest ski flight.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:12, the ski jumper begins to celebrate.", "inference": "Mark the point about the question that needs attention."}, {"modality": "vision", "evidence": "At 0:13, the ski jumper's team begins to celebrate.", "inference": "Mark the point about the question that needs attention."}, {"modality": "audio", "evidence": "At 0:20, the commentator says 'That is a world record.'", "inference": "They celebrate for the new world record."}]}
{"id": "video_626__q0000", "video_id": "video_626", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_626_t5s.mp4"], "extra_info": {"question": "Before the first modal particle appears, how many times does the central character in sunglasses raise their arm in sync with the vocalization of 'All My Fellas'?"}, "reward_model_ground_truth": {"multi_choice": ["A. 10 times", "B. 12 times", "C. 8 times", "D. 6 times"], "correct_option": "D"}, "answer_text": "6 times", "reasoning_steps": [{"modality": "audio", "evidence": "The first modal particle appears at 0:28.", "inference": "Count the character's arm raises before this time."}, {"modality": "vision", "evidence": "The central character in sunglasses appeared at 0:03.", "inference": "Pay attention to the character's moves next."}, {"modality": "vision", "evidence": "At 0:06, the character raised the arm in sync with the lyric subtitle 'All My Fellas'.", "inference": "Counted as 1 time so far."}, {"modality": "vision", "evidence": "At 0:08, the character raised the arm in sync with the lyric subtitle 'All My Fellas'.", "inference": "Counted as 2 times so far."}, {"modality": "vision", "evidence": "At 0:12, the character raised the arm in sync with the lyric subtitle 'All My Fellas'.", "inference": "Counted as 3 times so far."}, {"modality": "vision", "evidence": "At 0:15, the character raised the arm in sync with the lyric subtitle 'All My Fellas'.", "inference": "Counted as 4 times so far."}, {"modality": "vision", "evidence": "At 0:17, the character raised the arm in sync with the lyric subtitle 'All My Fellas'.", "inference": "Counted as 5 times so far."}, {"modality": "vision", "evidence": "At 0:20, the character raised the arm in sync with the lyric subtitle 'All My Fellas'.", "inference": "Counted as 6 times so far."}]}
{"id": "video_627__q0000", "video_id": "video_627", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_627_t5s.mp4"], "extra_info": {"question": "In which scene did the dog let out a wronged whimper?"}, "reward_model_ground_truth": {"multi_choice": ["A. The 7th stage", "B. The 8th stage", "C. The 9th stage", "D. The 6th stage"], "correct_option": "B"}, "answer_text": "The 8th stage", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:01 in the video, the text 'Player06 was last online 37 Days' was displayed.", "inference": "According to the question, this is the start of the first stage of time progression in the video."}, {"modality": "vision", "evidence": "At 0:02, the number '37' changed to '72'.", "inference": "This is the start of second stage."}, {"modality": "vision", "evidence": "At 0:04 in the video, the number '72' changed to '131'.", "inference": "This is the start of third stage."}, {"modality": "vision", "evidence": "At 0:06 in the video, the number '131' changed to '187'.", "inference": "This is the start of fourth stage."}, {"modality": "vision", "evidence": "At 0:08 in the video, the number '187' changed to '213'.", "inference": "This is the start of fifth stage."}, {"modality": "vision", "evidence": "At 0:09 in the video, the number '213' changed to '286'.", "inference": "This is the start of sixth stage."}, {"modality": "vision", "evidence": "At 0:11 in the video, the text changed to 'Player06 was last online 1 Year'.", "inference": "This is the start of seventh stage."}, {"modality": "vision", "evidence": "At 0:14 in the video, the text changed to 'Player06 was last online 2 Year'.", "inference": "This is the 8th stage of time progression."}, {"modality": "audio", "evidence": "At the 8th stage, the dog let out a wronged whimper.", "inference": "It is the 8th stage that the dog let out a wronged whimper."}]}
{"id": "video_628__q0000", "video_id": "video_628", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_628_t5s.mp4"], "extra_info": {"question": "When mentioning shorts and shirts, among the evaluations at the following different stages, which one is correct?"}, "reward_model_ground_truth": {"multi_choice": ["A.The expected evaluation of oversized shorts compared to striped shirts is more certain.", "B.The cute evaluation of the oversized white shirt is more prominent in tone compared to the previous one.", "C.The tone of the evaluation of the oversize blue coat is more excited than that of others.", "D.The tone of the evaluation of oversized shorts is more excited than that of shirts."], "correct_option": "B"}, "answer_text": "The cute evaluation of the oversized white shirt is more prominent in tone compared to the previous one.", "reasoning_steps": [{"modality": "vision", "evidence": "At 11:58, in the picture, the girl is holding a oversized white shirt in her hand.", "inference": "identify the oversized white shirt."}, {"modality": "vision", "evidence": "At 11:39, in the picture, the girl is holding a striped shirt in her hand.", "inference": "identify the previous clothing is a striped shirt."}, {"modality": "audio", "evidence": "At 11:39, the girl said, 'I have this striped one that is oversized and it's super cute......that's very convenient.'", "inference": "she thinks the striped shirt is cute and convenient."}, {"modality": "audio", "evidence": "At 11:58, when the girl said 'it's oversized and cute', she stretched out her voice to say the word 'cute'.", "inference": "she was excited about the oversized white shirt."}, {"modality": "vision", "evidence": "At 11:58,  the girl accompanied her voice with exaggerated and adorable hand gestures.", "inference": "she thinks the oversized white shirt is more prominent compared to the previous striped shirt."}]}
{"id": "video_628__q0001", "video_id": "video_628", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_628_t5s.mp4"], "extra_info": {"question": "When packing, what is the tone of the evaluation of the clothes that are being pressed under the colorful clothes to the left of the pink pouch?"}, "reward_model_ground_truth": {"multi_choice": ["A.Uncertain tone", "B.Calm tone", "C.Excited tone", "D.Disgusted tone"], "correct_option": "B"}, "answer_text": "Calm tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 18:33, in the picture, a pair of camo capris was pressed under the colorful clothes to the left of the pink pouch.", "inference": "the clothes that being pressed under the colorful clothes include a pair of camo capris."}, {"modality": "vision", "evidence": "At 18:35, in the picture, another pair of trousers can be seen after the camo capris was packed.", "inference": "the clothes that being pressed under the colorful clothes include the camo capris and another pair of trousers."}, {"modality": "audio", "evidence": "At 10:38, the girl said, 'then I have these camo capris that I recently thrifted, I think they are so cool.'", "inference": "she delivered the evaluation of these clothes in a 'Calm tone'."}]}
{"id": "video_628__q0002", "video_id": "video_628", "qid": 2, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_628_t5s.mp4"], "extra_info": {"question": "what was the girl's tone like when introducing the oversized shirt?"}, "reward_model_ground_truth": {"multi_choice": ["A.a doubtful tone", "B.an excited tone", "C.a cherished tone", "D.a regretful tone"], "correct_option": "B"}, "answer_text": "an excited tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 11:58, in the picture, the girl is holding a oversized white shirt in her hand.", "inference": "identify the oversized shirt."}, {"modality": "audio", "evidence": "At 11:58, the girl said, 'it's oversized and cute.'", "inference": "she thinks oversized shirt is cute."}, {"modality": "audio", "evidence": "At 11:58, when the girl said 'it's oversized and cute', she stretched out her voice to say the word 'cute'.", "inference": "she was excited about it."}]}
{"id": "video_63__q0000", "video_id": "video_63", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_63_t5s.mp4"], "extra_info": {"question": "When the blogger mentioned 'This entire community survives on water collected from puddles like this.', which color of bucket is the farthest from the woman in the picture?"}, "reward_model_ground_truth": {"multi_choice": ["A.White", "B.Orange", "C.Purple", "D.Grey"], "correct_option": "D"}, "answer_text": "Grey", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:56, the blogger mentioned, 'this entire community survives on water collected from puddles like this.'", "inference": "the blogger mentioned the puddles like the one in the vision"}, {"modality": "vision", "evidence": "At 0:56, there are many buckets beside the woman.", "inference": "there are many buckets beside the woman in the picture."}, {"modality": "vision", "evidence": "At 0:56, the grey bucket is the farthest from the woman in the picture.", "inference": "the grey bucket is the farthest from the woman in the picture."}]}
{"id": "video_63__q0001", "video_id": "video_63", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_63_t5s.mp4"], "extra_info": {"question": "In the drinking water filtration system built by the blogger in Colombia, how many tanks are there in total for the parts responsible for filtering sand and gravel and removing salt?"}, "reward_model_ground_truth": {"multi_choice": ["A.3.", "B.4.", "C.5.", "D.7."], "correct_option": "D"}, "answer_text": "7", "reasoning_steps": [{"modality": "vision", "evidence": "At 2:36, there are 3 big blue tanks in the picture.", "inference": "there are 3 tanks in the first stage of the system."}, {"modality": "audio", "evidence": "At 2:36, the blogger said 'this removes dirt sand and large particles'.", "inference": "there are 3 tanks for the part responsible for filtering sand."}, {"modality": "vision", "evidence": "At 2:41, there are 4 white tanks in the picture.", "inference": "there are 4 tanks in the third stage of the system."}, {"modality": "audio", "evidence": "At 2:41, the blogger said 'and then is pumped into here which removes the saline'.", "inference": "there are 4 tanks for the part responsible for removing salt. So in total, there are 7 tanks."}]}
{"id": "video_63__q0002", "video_id": "video_63", "qid": 2, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_63_t5s.mp4"], "extra_info": {"question": "When Teamseas appears in the video, who is introducing it?"}, "reward_model_ground_truth": {"multi_choice": ["A.Mark Rober.", "B.Ben Azelart.", "C.Stokes Twins.", "D.The blogger."], "correct_option": "D"}, "answer_text": "The blogger", "reasoning_steps": [{"modality": "vision", "evidence": "At 9:13, Teamseas appears.", "inference": "Teamseas appears."}, {"modality": "audio", "evidence": "At 9:13, a man's voice is introducing Teamseas.", "inference": "a man's voice is introducing Teamseas."}, {"modality": "audio", "evidence": "from the entire video, it can be easily known that the man's voice introducing Teamseas belongs to the blogger.", "inference": "the blogger is introducing Teamseas."}]}
{"id": "video_64__q0000", "video_id": "video_64", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_64_t5s.mp4"], "extra_info": {"question": "What was the blogger doing when the alarm went off?"}, "reward_model_ground_truth": {"multi_choice": ["A.Running.", "B.Tug of war.", "C.Eating chocolate.", "D.Riding a bicycle."], "correct_option": "A"}, "answer_text": "Running.", "reasoning_steps": [{"modality": "audio", "evidence": "At 9:22, the alarm sound rang in the background sound.", "inference": "This is the moment mentioned in the question."}, {"modality": "vision", "evidence": "At 9:22, the blogger appeared in the scene.", "inference": "Pay attention to what he was doing in at the moment."}, {"modality": "vision", "evidence": "At 9:22, the blogger was running.", "inference": "So the answer is 'Running.'."}]}
{"id": "video_65__q0000", "video_id": "video_65", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_65_t5s.mp4"], "extra_info": {"question": "What was the second skill used in the last short round before the blogger said 'quick before this video continues'?"}, "reward_model_ground_truth": {"multi_choice": ["A.The first skill at the bottom of the screen.", "B.The second skill at the bottom of the screen.", "C.The third skill at the bottom of the screen.", "D.The fourth skill at the bottom of the screen."], "correct_option": "B"}, "answer_text": "The second skill at the bottom of the screen.", "reasoning_steps": [{"modality": "audio", "evidence": "At 1:02, the blogger said'quick before this video continues'.", "inference": "Here is the time point."}, {"modality": "vision", "evidence": "At 0:47, the first skill used is the third skill at the bottom of the screen.", "inference": "The first skill is the third skill at the bottom of the screen."}, {"modality": "vision", "evidence": "At 0:52, the second skill used is the second at the bottom of the screen.", "inference": "The second skill is the second skill at the bottom of the screen."}]}
{"id": "video_66__q0000", "video_id": "video_66", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_66_t5s.mp4"], "extra_info": {"question": "Before the blogger said that the RE-45 fires very quickly, how much damage did he deal to the opponent with the RE-45?"}, "reward_model_ground_truth": {"multi_choice": ["A.53.", "B.81.", "C.28.", "D.134."], "correct_option": "D"}, "answer_text": "134.", "reasoning_steps": [{"modality": "audio", "evidence": "The blogger said he got the RE-45 at 11:20.", "inference": "The RE-45 is mentioned in the question."}, {"modality": "audio", "evidence": "The blogger said this thing (RE-45) fired so fast at 11:26.", "inference": "The question is asking how much damage the blogger dealt to the opponent with the RE-45 before 11:26."}, {"modality": "vision", "evidence": "At 11:24, the points of damage displayed in the center of the video screen was 53.", "inference": "He first dealt 53 points of damage at 11:24."}, {"modality": "vision", "evidence": "At 11:25, the points of damage displayed in the center of the video screen was 81.", "inference": "He then dealt 81 points of damage at 11:25. Based on the above points of damage, we can do the calculation: 53+81=134. So the blogger dealt 134 points of damage to the opponent with the RE-45."}]}
{"id": "video_66__q0001", "video_id": "video_66", "qid": 1, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_66_t5s.mp4"], "extra_info": {"question": "How many bullets did the blogger fire when saying 'win my first one of the season'?"}, "reward_model_ground_truth": {"multi_choice": ["A.15.", "B.24.", "C.240.", "D.9."], "correct_option": "A"}, "answer_text": "15.", "reasoning_steps": [{"modality": "audio", "evidence": "The blogger said 'win my first one of the season' at 1:25 in the video.", "inference": "The question is asking for the number of bullets the blogger fired at this approximate time."}, {"modality": "vision", "evidence": "The blogger got the Hemlok (rifle) at 1:17.", "inference": "We know when he first got the rifle."}, {"modality": "vision", "evidence": "At 1:18, the total number of bullets displayed at the bottom right corner of the video was 24.", "inference": "The rifle had 24 bullets in total."}, {"modality": "vision", "evidence": "At 1:20, the blogger continued to shoot at an opponent.", "inference": "The blogger was shooting."}, {"modality": "vision", "evidence": "At 1:22, the number of bullets displayed at the bottom right corner of the video was 9.", "inference": "Based on the above inferences, we can do the calculation: 24-9=15. So the blogger fired 15 bullets when saying 'win my first one of the season'."}]}
{"id": "video_66__q0002", "video_id": "video_66", "qid": 2, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_66_t5s.mp4"], "extra_info": {"question": "Before the blogger said 'what's up buddy you're dead' for the first time, how many bullets did his character use to kill the opponent?"}, "reward_model_ground_truth": {"multi_choice": ["A.15.", "B.24.", "C.18.", "D.9."], "correct_option": "C"}, "answer_text": "18.", "reasoning_steps": [{"modality": "audio", "evidence": "The blogger said 'what's up buddy you're dead' at 1:42 in the video.", "inference": "The question is asking for the number of bullets the blogger used to kill the opponent at this approximate time."}, {"modality": "vision", "evidence": "At 1:40, the number of bullets displayed at the bottom right corner of the video was 23.", "inference": "The total bullets number was 23 at that moment."}, {"modality": "vision", "evidence": "At 1:41, the blogger continuously fired at the opponent.", "inference": "The blogger used bullets to kill the opponent."}, {"modality": "vision", "evidence": "At 1:42, the number of bullets displayed at the bottom right corner of the video was 5.", "inference": "Based on the above inferences, we can do the calculation: 23-5=18, so the blogger used 18 bullets to kill the opponent."}]}
{"id": "video_67__q0000", "video_id": "video_67", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_67_t5s.mp4"], "extra_info": {"question": "What did the person with the same beard as the blogger eat?"}, "reward_model_ground_truth": {"multi_choice": ["A.Steak.", "B.Ice cream.", "C.Hamburger.", "D.Mushroom."], "correct_option": "C"}, "answer_text": "Hamburger.", "reasoning_steps": [{"modality": "vision", "evidence": "At the timestamp 3:05,there was a person with the same beard as the blogger.", "inference": "We need to find out what did he eat."}, {"modality": "audio", "evidence": "At the timestamp 3:44, he said that he had a double burger.", "inference": "He ate a hamburger."}]}
{"id": "video_67__q0001", "video_id": "video_67", "qid": 1, "question_type": "spatial understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_67_t5s.mp4"], "extra_info": {"question": "When the music 'The Sound of Silence' starts playing, what color is the clothes worn by the man in the middle of the frame?"}, "reward_model_ground_truth": {"multi_choice": ["A.White.", "B.Black.", "C.Grey.", "D.Red."], "correct_option": "A"}, "answer_text": "White.", "reasoning_steps": [{"modality": "audio", "evidence": "At the timestamp 5:28，the music The Sound of Silence starts playing.", "inference": "Here is the timepoiont."}, {"modality": "vision", "evidence": "The man appears in the shot is wearing a white cloth.", "inference": "The color is white."}]}
{"id": "video_67__q0002", "video_id": "video_67", "qid": 2, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_67_t5s.mp4"], "extra_info": {"question": "When the music The Sound of Silence starts playing, which of the following items has never appeared on the table of the man in the center of the frame?"}, "reward_model_ground_truth": {"multi_choice": ["A.Tomato sauce.", "B.Steak.", "C.Dessert.", "D.Beer glass."], "correct_option": "B"}, "answer_text": "Steak.", "reasoning_steps": [{"modality": "audio", "evidence": "At the timestamp 5:28，the music plays.", "inference": "Pay attention to the frame about things appeared."}, {"modality": "vision", "evidence": "There are tomato sauce and beer glass in front of him.", "inference": "Tomato sauce and beer glass are appeared on the table."}, {"modality": "audio", "evidence": "At the timestamp 12:00, the bloggers says 'that's a cheese cake?'", "inference": "He wants to know what a dessert is that."}, {"modality": "vision", "evidence": "The dessert is in front of him.", "inference": "So the steak has never appeared on the table of the man in the center of the frame.."}]}
{"id": "video_68__q0000", "video_id": "video_68", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_68_t5s.mp4"], "extra_info": {"question": "After the anchor said 'being high ranking in this game seems so punishing', what is the key for the second skill he released?"}, "reward_model_ground_truth": {"multi_choice": ["A.Q.", "B.E.", "C.Lshift.", "D.Right mouse button."], "correct_option": "D"}, "answer_text": "Right mouse button.", "reasoning_steps": [{"modality": "audio", "evidence": "The anchor said 'being high ranking in this game seems so punishing' at 3:16.", "inference": "To answer the question, we need to focus on the second skill released after 3:16."}, {"modality": "vision", "evidence": "The first skill released was the first one in the bottom right corner at 3:22.", "inference": "This is the first skill he released."}, {"modality": "vision", "evidence": "The second skill released was the third one in the bottom right corner at 3:25.", "inference": "This is the second skill he released."}, {"modality": "vision", "evidence": "At 3:25, we can see the corresponding key for the third skill in the bottom right corner is right mouse button.", "inference": "After the anchor said 'being high ranking in this game seems so punishing', the key for the second skill he released is right mouse button."}]}
{"id": "video_69__q0000", "video_id": "video_69", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_69_t5s.mp4"], "extra_info": {"question": "When the blogger wanted a large container to hold yogurt, where was the yogurt on his right side when he started eating?"}, "reward_model_ground_truth": {"multi_choice": ["A.At the closest position to his left.", "B.At the second closest position to his left.", "C.At the farthest position to his left.", "D.To his right."], "correct_option": "A"}, "answer_text": "At the closest position to his left.", "reasoning_steps": [{"modality": "audio", "evidence": "he mentioned it at 6:33.", "inference": "locate it."}, {"modality": "vision", "evidence": "the yogurt is on his right side.", "inference": "identify it."}, {"modality": "vision", "evidence": "he started to eat at 3:47.", "inference": "there are 3 yogurt to his left."}, {"modality": "vision", "evidence": "he picked up the closest yogurt on his left side at 4:56.", "inference": "get the point."}, {"modality": "vision", "evidence": "he put the yogurt on his right side at 5:05.", "inference": "At the closest position to his left."}]}
{"id": "video_69__q0001", "video_id": "video_69", "qid": 1, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_69_t5s.mp4"], "extra_info": {"question": "When the blogger wanted a large container to hold yogurt, where was the closest yogurt on his left side when he started eating?"}, "reward_model_ground_truth": {"multi_choice": ["A.At the closest position to his left.", "B.At the second closest position to his left.", "C.At the farthest position to his left.", "D.To his right."], "correct_option": "B"}, "answer_text": "At the second closest position to his left.", "reasoning_steps": [{"modality": "audio", "evidence": "he mentioned it at 6:33.", "inference": "locate it."}, {"modality": "vision", "evidence": "he started to eat at 3:35.", "inference": "identify the closest yogurt on his left side."}, {"modality": "vision", "evidence": "he picked up the closest yogurt on his left side at 4:56.", "inference": "get the point."}, {"modality": "vision", "evidence": "he put the yogurt on his right side at 5:05.", "inference": "the yogurt at the second closest position to his left side changed to the closest position to his left."}]}
{"id": "video_7__q0000", "video_id": "video_7", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_7_t5s.mp4"], "extra_info": {"question": "What is Josh Allen's role in the team?"}, "reward_model_ground_truth": {"multi_choice": ["A. He is primarily focused on individual performance and skill development, with limited engagement in team-building or mentorship roles", "B. He serves as a central figure in fostering team camaraderie and maintaining a positive team culture, actively engaging with players at all levels", "C. He primarily interacts with senior players and coaches, suggesting a hierarchical structure that limits his involvement with newer or less experienced teammates.", "D. His interactions are largely confined to on-field activities, suggesting a separation between his professional role and personal relationships within the team."], "correct_option": "B"}, "answer_text": "He serves as a central figure in fostering team camaraderie and maintaining a positive team culture, actively engaging with players at all levels", "reasoning_steps": [{"modality": "vision", "evidence": "the man in red, who wears the '17' on the clothes is Allen.", "inference": "get the person."}, {"modality": "audio", "evidence": "the man always said 'oh nice'.", "inference": "he always maintains positive."}, {"modality": "vision", "evidence": "he played with every players.", "inference": "he serves as a central figure in fostering team camaraderie and maintaining a positive team culture, actively engaging with players at all levels"}]}
{"id": "video_70__q0000", "video_id": "video_70", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_70_t5s.mp4"], "extra_info": {"question": "Who asked the protagonist to remain silent?"}, "reward_model_ground_truth": {"multi_choice": ["A. Eleanor Twitty", "B. Spider Witch", "C. Ray", "D. Slimer"], "correct_option": "A"}, "answer_text": "Eleanor Twitty", "reasoning_steps": [{"modality": "vision", "evidence": "At 5:57, the scene shows that the protagonist is in a library.", "inference": "The library is a place where one needs to keep quiet."}, {"modality": "audio", "evidence": "At 6:00, The protagonist said loudly 'Eleanor Twitty! The librarian!'.", "inference": "At this moment, the protagonist excitedly identify a person."}, {"modality": "audio", "evidence": "At 6:01, a woman's shush can be heard.", "inference": "She is reminding the protagonist to keep quiet."}, {"modality": "vision", "evidence": "At 6:01, a ghost - like female librarian appears in the picture..", "inference": "She is the one who makes the protagonist keep quiet, 'Eleanor Twitty'."}]}
{"id": "video_70__q0001", "video_id": "video_70", "qid": 1, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_70_t5s.mp4"], "extra_info": {"question": "Who keeps making noise when the roller coaster is flying through the air?"}, "reward_model_ground_truth": {"multi_choice": ["A. Marshmallow Man", "B. Slimer", "C. Spider Witch", "D. Sherman Twins"], "correct_option": "D"}, "answer_text": "Sherman Twins", "reasoning_steps": [{"modality": "vision", "evidence": "At 8:23, the roller coaster is flying through the air in the scene.", "inference": "This is the moment mentioned in the question."}, {"modality": "vision", "evidence": "At 8:23, Sherman twins(women) are flying around the roller coaster in the scene.", "inference": "They may be the person who keeps making noise."}, {"modality": "audio", "evidence": "From 8:23, The murmurs of women can be heard constantly in the background sound.", "inference": "So this is the 'Sherman Twins' keeps making noise."}]}
{"id": "video_71__q0000", "video_id": "video_71", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_71_t5s.mp4"], "extra_info": {"question": "What were my emotions like when I initially got in the car?"}, "reward_model_ground_truth": {"multi_choice": ["A. Easiness.", "B. Shock.", "C. Fear.", "D. Joyance."], "correct_option": "B"}, "answer_text": "Shock.", "reasoning_steps": [{"modality": "vision", "evidence": "At 00:29, the man opened the car door.", "inference": "the man was about to get into the car."}, {"modality": "vision", "evidence": "At 00:29, there is a dirty patch on the road beside the car.", "inference": "it's dirty."}, {"modality": "audio", "evidence": "At 00:29, the man said 'what the hell is that'.", "inference": "the man was shocked."}]}
{"id": "video_71__q0001", "video_id": "video_71", "qid": 1, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_71_t5s.mp4"], "extra_info": {"question": "What things did I do between purchasing the items in the second order and delivering those items?"}, "reward_model_ground_truth": {"multi_choice": ["A. Rejecting 2 orders.", "B. Introducing the red card.", "C. Fueling the car.", "D. Complaining about the car."], "correct_option": "B"}, "answer_text": "Introducing the red card.", "reasoning_steps": [{"modality": "audio", "evidence": "At 03:25, the man said, 'ten dollars and 50 cents point nine miles for 17 items, so I'm gonna turn the car off, because that is a lot of items.'", "inference": "the man accepted this order and going to purchasing the items."}, {"modality": "vision", "evidence": "At 04:46, the man pushed the shopping cart to the side of the vehicle.", "inference": "he purchased the items."}, {"modality": "audio", "evidence": "At 04:47, the man said  'yeah two bag, and three bags'.", "inference": "the man was introducing the red card."}, {"modality": "audio", "evidence": "At 05:04, the man said '3 minute drive to the drop off.'", "inference": "the man was going to delivering those items. So between purchasing the items in the second order and delivering those items, the man was introducing the red card."}]}
{"id": "video_74__q0000", "video_id": "video_74", "qid": 0, "question_type": "counting", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_74_t5s.mp4"], "extra_info": {"question": "How many children are on the sandy ground just as the whistle blows?"}, "reward_model_ground_truth": {"multi_choice": ["A. 1.", "B. 2.", "C. 3.", "D. 4."], "correct_option": "D"}, "answer_text": "4.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:04, the whistle blows.", "inference": "Here is the time point."}, {"modality": "vision", "evidence": "There are four children are on the sandy ground.", "inference": "There are four children are on the sandy ground just as the whistle blows"}]}
{"id": "video_74__q0001", "video_id": "video_74", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_74_t5s.mp4"], "extra_info": {"question": "What drink is in the mother's hand when the boy is laughing with her?"}, "reward_model_ground_truth": {"multi_choice": ["A. Coffee.", "B. Cereal milk.", "C. Orange juice.", "D. Milk."], "correct_option": "C"}, "answer_text": "Orange juice.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:56, the boy is laughing with his mother.", "inference": "Here is the time point."}, {"modality": "vision", "evidence": "the mother is holing a carton and standing near the boy.", "inference": "Notice the drink pouring from the carton."}, {"modality": "vision", "evidence": "At 1:03, the carton with name of OJ.", "inference": "The drink is Orange juice."}]}
{"id": "video_74__q0002", "video_id": "video_74", "qid": 2, "question_type": "spatial understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_74_t5s.mp4"], "extra_info": {"question": "When the teacher in purple coughs, what is the third letter on the blackboard?"}, "reward_model_ground_truth": {"multi_choice": ["A. A.", "B. B.", "C. C.", "D. D."], "correct_option": "D"}, "answer_text": "D.", "reasoning_steps": [{"modality": "audio", "evidence": "At 3:50, a person coughs.", "inference": "Here is the time point."}, {"modality": "vision", "evidence": "The teacher in purple coughs.", "inference": "The blackboard is behind the teacher."}, {"modality": "vision", "evidence": "There are many letters on the blackboard.", "inference": "The third letter is D."}]}
{"id": "video_76__q0000", "video_id": "video_76", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_76_t5s.mp4"], "extra_info": {"question": "Which clone comic has won the first place in the protagonist's mind?"}, "reward_model_ground_truth": {"multi_choice": ["A. The person in blue.", "B. The person in yellow.", "C. The person in grey.", "D. The person in purple."], "correct_option": "D"}, "answer_text": "The person in purple.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:57, a man said' clone..'.", "inference": "locate the 'clone'."}, {"modality": "vision", "evidence": "they show many clone comics.", "inference": "pay attention to them."}, {"modality": "audio", "evidence": "At 8:01,a man said ' it's good..'.", "inference": "maybe it is a good comic."}, {"modality": "vision", "evidence": "the comic is showed by the one in purple.", "inference": "The person in purple."}]}
{"id": "video_77__q0000", "video_id": "video_77", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_77_t5s.mp4"], "extra_info": {"question": "When the blogger was marinating the chicken, on which side of the 6th seasoning was the 5th seasoning from the blogger's perspective?"}, "reward_model_ground_truth": {"multi_choice": ["A.Front", "B.Back", "C.Left", "D.Right"], "correct_option": "D"}, "answer_text": "Right", "reasoning_steps": [{"modality": "vision", "evidence": "At 3:08, the blogger was preparing the marinade for chicken.", "inference": "It indicates that he would use various seasonings afterward."}, {"modality": "audio", "evidence": "At 3:10, the blogger said 'Vinegar'.", "inference": "So the first seasoning is vinegar."}, {"modality": "audio", "evidence": "At 3:11, the blogger said 'soy sauce'.", "inference": "So the second seasoning is soy sauce."}, {"modality": "audio", "evidence": "At 3:13, the blogger said 'black pepper'.", "inference": "So the third seasoning is black pepper."}, {"modality": "audio", "evidence": "At 3:14, the blogger said 'chili flakes'.", "inference": "So the 4th seasoning is chili flakes."}, {"modality": "audio", "evidence": "At 3:15, the blogger said 'a couple of bay leaves'.", "inference": "So the 5th seasoning is bay leaves."}, {"modality": "audio", "evidence": "At 3:17, the blogger said 'just a little bit of salt'.", "inference": "So the 6th seasoning is salt."}, {"modality": "vision", "evidence": "At 2:46, all the seasonings are presented on the table.", "inference": "Identify the position of the bay leaves for the salt."}, {"modality": "vision", "evidence": "At 2:46, the bay leaves are to the right of the salt.", "inference": "So the answer is 'Right'."}]}
{"id": "video_77__q0001", "video_id": "video_77", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_77_t5s.mp4"], "extra_info": {"question": "Who is the person cutting the grapes?"}, "reward_model_ground_truth": {"multi_choice": ["A.Magnus Nilsson", "B.Michael Jiro Holman", "C.Edward Lee", "D.Josh Mayer"], "correct_option": "A"}, "answer_text": "Magnus Nilsson", "reasoning_steps": [{"modality": "audio", "evidence": "At 14:38, a man said 'I'm just gonna cut two thin slices of that.'.", "inference": "This indicates that the man is the person who is cutting the grapes into slices, identify him."}, {"modality": "vision", "evidence": "At 14:38, a long - haired man appears in the scene alone.", "inference": "Pay attention to his name."}, {"modality": "vision", "evidence": "At 12:23, this long - haired man appears in the first time.", "inference": "An introduction about him may emerge in the scene."}, {"modality": "vision", "evidence": "At 12:23, a line of text 'CHEF MAGNUS NILSSON' displayed at the bottom of the scene.", "inference": "The long - haired man's name is 'Magnus Nilsson'."}]}
{"id": "video_77__q0002", "video_id": "video_77", "qid": 2, "question_type": "reference reasoning", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_77_t5s.mp4"], "extra_info": {"question": "Who commented after eating the steak?"}, "reward_model_ground_truth": {"multi_choice": ["A.Magnus Nilsson", "B.Michael Jiro Holman", "C.Fred Morin", "D.David Chang"], "correct_option": "D"}, "answer_text": "David Chang", "reasoning_steps": [{"modality": "audio", "evidence": "At 17:47, a man with a beard commentes Joe BEEF when it was being made.", "inference": "Pay attention to his name."}, {"modality": "vision", "evidence": "At 15:35, this bearded man appeared in the first time.", "inference": "An introduction about him may emerge in the scene."}, {"modality": "vision", "evidence": "At 15:34, a line of text 'CHEF FRED MORIN' displayed at the bottom of the scene.", "inference": "The bearded man's name is 'FRED MORIN'."}]}
{"id": "video_78__q0000", "video_id": "video_78", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_78_t5s.mp4"], "extra_info": {"question": "What is the maximum number of lemons placed on the table in total?"}, "reward_model_ground_truth": {"multi_choice": ["A.3", "B.5", "C.8", "D.11"], "correct_option": "D"}, "answer_text": "11", "reasoning_steps": [{"modality": "audio", "evidence": "At 05:46, the man said, 'now let's make the lemonade with all the lemons we got'.", "inference": "identify all the lemons placed on the table."}, {"modality": "vision", "evidence": "At 06:18, there are 10 whole lemons on the table.", "inference": "identify 10 whole lemons on the table."}, {"modality": "vision", "evidence": "At 06:18, there is one sliced lemon on the cutting board.", "inference": "identify one sliced lemon on the cutting board. so there are eleven lemons in total."}]}
{"id": "video_79__q0000", "video_id": "video_79", "qid": 0, "question_type": "spatial understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_79_t5s.mp4"], "extra_info": {"question": "When the woman is packing, which brand/store is the colorful dress to the left of the pink pouch from?"}, "reward_model_ground_truth": {"multi_choice": ["A.other stories", "B.Bash Paris", "C.Vera Moda", "D.Uniqlo"], "correct_option": "B"}, "answer_text": "Bash Paris", "reasoning_steps": [{"modality": "vision", "evidence": "From 17:29 of the video, the woman is packing, there is a colorful dress to the left of the pink pouch.", "inference": "Confirm the colorful dress first."}, {"modality": "audio", "evidence": "From 13:25 of the video,she introduced the dress is from 'Bash Paris'.", "inference": "Bash Paris is the brand of the colorful dress to the left of the pink pouch."}]}
{"id": "video_79__q0001", "video_id": "video_79", "qid": 1, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_79_t5s.mp4"], "extra_info": {"question": "Regarding the sequence of the following video content, which one is correct?"}, "reward_model_ground_truth": {"multi_choice": ["A.Introduce the overall plan->Checking the purchased items->Shopping for a power bank while wandering the streets->Ticket accident", "B.Shopping for a power bank while wandering the streets->Shopping at the camera store->Ticket accident->Organizing daily necessities", "C.Shopping at the camera store->Checking the purchased items->Ticket accident->Organizing daily necessities->Eating", "D.Introduce the overall plan->Shopping for a power bank while wandering the streets->Eating->Ticket accident"], "correct_option": "C"}, "answer_text": "Introduce the overall plan->Shopping for a power bank while wandering the streets->Organizing daily necessities->Ticket accident", "reasoning_steps": [{"modality": "audio", "evidence": "From  the beginning of the video, the female lead begins with the overall plan.", "inference": "The first step is introducing the overall plan."}, {"modality": "audio", "evidence": "From 3:49 of the video, she says'I have arrived at the camera store'", "inference": "From now on, she was going out for shopping."}, {"modality": "vision", "evidence": "From 4:11 of the video, she bought a power bank while wandering the streets.", "inference": "The second step is shopping for a power bank while wandering the streets."}, {"modality": "audio", "evidence": "From 5:47 of the video, she said:'we are going to pack everything'.", "inference": "The third step is organizing daily necessities."}, {"modality": "audio", "evidence": "From 15:48 of the video, she said'the train we are planning to take tomorrow is fully booked '", "inference": "There is a ticket accident happened."}]}
{"id": "video_79__q0002", "video_id": "video_79", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_79_t5s.mp4"], "extra_info": {"question": "When the girl starts introducing the clothes she is going to take, which of the following is the correct order of descriptions for the clothes?"}, "reward_model_ground_truth": {"multi_choice": ["A.using quite a lot , so cool for orange ,so cool, super elegant ", "B.oversize but super cute, matching for striped one, mathcing for shorts ,black for save space  ", "C.oversize but super cute, mathcing for shorts , matching for striped one, black for save space  ", "D.using quite a lot , so cool for orange ,super elegant , pretty qute"], "correct_option": "D"}, "answer_text": "using quite a lot , so cool for orange ,quite elegant for looking like skinny, pretty cute", "reasoning_steps": [{"modality": "vision", "evidence": "At 10:30, she showed us a pair of denim shorts.", "inference": "This is the first cloth the girl introduced to us."}, {"modality": "audio", "evidence": "She said,'that I think I'll be using quite a lot.'.", "inference": "The first description for the clothes is 'using quite a lot'."}, {"modality": "vision", "evidence": "From 10:41 of the video, she showed us a pair of orange capris.", "inference": "This is the next clothe the girl introduced to us."}, {"modality": "audio", "evidence": "She said,'they are so cool'", "inference": "Her next description is 'so cool for orange'"}, {"modality": "vision", "evidence": "From 10:51 of the video, she showed us a pair of another capris.", "inference": "This is the next clothe the girl introduced to us."}, {"modality": "audio", "evidence": "At 11:00, she said,'and they are very like skinny,super elegant'.", "inference": "Her next description is 'quite elegant for looking like skinny'."}, {"modality": "vision", "evidence": "From 11:14 of the video, she tried a pair of pants infront of the mirror.", "inference": "This is the next clothe the girl introduced to us."}, {"modality": "audio", "evidence": "She said,'I think it's pretty cute'.", "inference": "Her next description is 'pretty cute'."}]}
{"id": "video_79__q0003", "video_id": "video_79", "qid": 3, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_79_t5s.mp4"], "extra_info": {"question": "What did they do on the second day of preparing for the trip?"}, "reward_model_ground_truth": {"multi_choice": ["A.They went for shopping.", "B.They managed to book the train tickets.", "C.They booked Airbnb", "D.Both A and C."], "correct_option": "C"}, "answer_text": "They booked Airbnb", "reasoning_steps": [{"modality": "audio", "evidence": "From 5:06 of the video, the girl says 'hello again, it is actually the day after...'.", "inference": "Things happened in this clip is on the second day."}, {"modality": "vision", "evidence": "From 14:30 of the video, the girl shows her computer with Airbnb.", "inference": "She was using Airbnb on the second day."}, {"modality": "audio", "evidence": "From 14:30 of the video, the girl said:'Because we finally booked an Airbnb in Berlin.'", "inference": "They booked Airbnb on the second day of preparing for the trip."}]}
{"id": "video_8__q0000", "video_id": "video_8", "qid": 0, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_8_t5s.mp4"], "extra_info": {"question": "Why is Kenny dressed so distinctively at the game?"}, "reward_model_ground_truth": {"multi_choice": ["A. Because Kenny is assigned as a special correspondent.", "B. Because Kenny's outfit is a fashionable ensemble.", "C. Because Kenny himself is exactly a huge Taylor Swift fan.", "D. Because Kenny is trying to hide the hot dogs he ate during the game."], "correct_option": "A"}, "answer_text": "Because Kenny is assigned as a special correspondent.", "reasoning_steps": [{"modality": "audio", "evidence": "The host says' Let's go to our newest NFL sideline reporter Kenny Ditullio'. at 2:15.", "inference": "Kenny is assigned as the correspondent."}, {"modality": "vision", "evidence": "A man dresses so distinctively showing up in the scene at 2:23.", "inference": "This man is Kenny."}, {"modality": "audio", "evidence": "Then Kenny begins to report the game.", "inference": "Kenny dressed so distinctively because Kenny is assigned as a special correspondent."}]}
{"id": "video_80__q0000", "video_id": "video_80", "qid": 0, "question_type": "background&music understanding", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_80_t5s.mp4"], "extra_info": {"question": "What does the background music indicate?"}, "reward_model_ground_truth": {"multi_choice": ["A.He's pleased with the work that he is doing.", "B.He gets exhausted from his busy work schedule.", "C.He enjoys the short breaks during work.", "D.He feels hopeful about overcoming obstacles."], "correct_option": "A"}, "answer_text": "He's pleased with the work that he is doing.", "reasoning_steps": [{"modality": "audio", "evidence": "At 16:11, midway through the video, background music begins to play.", "inference": "Sets the point to further analyze the background music."}, {"modality": "audio", "evidence": "Then background music is bright.", "inference": "So the music might be indicating the man's cheerful mood."}, {"modality": "vision", "evidence": "At 16:54, Tte protagonist is loading peanuts into the trailer behind the tractor.", "inference": "The working is going smoothly and nicely."}, {"modality": "vision", "evidence": "At 16:45, he looks relax when working.", "inference": "He feels good about the work."}]}
{"id": "video_80__q0001", "video_id": "video_80", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_80_t5s.mp4"], "extra_info": {"question": "Which supply delivery required waiting for someone to deliver the supplies?"}, "reward_model_ground_truth": {"multi_choice": ["A.1st", "B.2nd", "C.3rd", "D.4th"], "correct_option": "B"}, "answer_text": "2nd receiving", "reasoning_steps": [{"modality": "vision", "evidence": "The first supplies receiving was a new seed tender at 02:06.", "inference": "Mark the 1st receiving."}, {"modality": "vision", "evidence": "The second  supplies receiving was a refill of fertilizer and hydrogen at 17:44.", "inference": "Mark the 2nd receiving."}, {"modality": "audio", "evidence": "At 17:44,The male protagonist said that they needed to wait for someone to deliver the supplies before they could go planting again.", "inference": "The second supply delivery required waiting for someone to deliver"}]}
{"id": "video_80__q0002", "video_id": "video_80", "qid": 2, "question_type": "hypothetical reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_80_t5s.mp4"], "extra_info": {"question": "What directly happens if the fuse blows?"}, "reward_model_ground_truth": {"multi_choice": ["A.Crops cannot be planted", "B.The circuit will be paralyzed", "C.Control panel' value displays as 0", "D.Peanuts can still be planted"], "correct_option": "C"}, "answer_text": "Control panel' value displays as 0", "reasoning_steps": [{"modality": "audio", "evidence": "The narrator states that the value displayed on the control panel indicates the pounds of pressure within the airbags.", "inference": "Exlain to audience how to observe the displaying information."}, {"modality": "vision", "evidence": "The video visually shows the control panel with a numerical reading.", "inference": "We should pay attention to these readings while working."}, {"modality": "audio", "evidence": "The narrator mentions that the target pressure for planting peanuts is 14.", "inference": "Confrim the target pressure level."}, {"modality": "audio", "evidence": "It is explained that a blown fuse will cause the air compressor to stop working.", "inference": "If the fuse blows, it would lead to the control panel's value displaying as 0."}]}
{"id": "video_80__q0003", "video_id": "video_80", "qid": 3, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_80_t5s.mp4"], "extra_info": {"question": "What was the protagonist's tone like when collaborating with others for the first time?"}, "reward_model_ground_truth": {"multi_choice": ["A.an excited tone", "B.a doubtful tone", "C.an awkward tone", "D.a regretful tone"], "correct_option": "C"}, "answer_text": "an awkward tone", "reasoning_steps": [{"modality": "vision", "evidence": "At 8:09, the protagonist was helping others pack peanuts into the box.", "inference": "This is the moment mentioned in the question,  pay attention to the tone of the protagonist."}, {"modality": "vision", "evidence": "At 8:15, the box had been filled with peanuts.", "inference": "This indicates that the switch for pouring peanuts should be turned off at this point."}, {"modality": "audio", "evidence": "At 8:15, others said excitedly, 'Turn it off!'", "inference": "The protagonist might not have turned off the switch in time."}, {"modality": "audio", "evidence": "At 8:18, the protagonist gave an awkward chuckle.", "inference": "This indicates that the protagonist felt embarrassed because he didn't turn off the switch in time, so the answer is 'an awkward tone'."}]}
{"id": "video_80__q0004", "video_id": "video_80", "qid": 4, "question_type": "causal reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_80_t5s.mp4"], "extra_info": {"question": "After getting off the tractor for the first time, why the man's voice increases suddenly?"}, "reward_model_ground_truth": {"multi_choice": ["A.Because he is very pleased with his peanut harvest.", "B.Because he is in such a hurry that some peanuts wastes.", "C.Because he thinks his companion's way of dealing with the peanut is perfect.", "D.Because he wants to get back to the tractor as quickly as possible."], "correct_option": "B"}, "answer_text": "Because he is in such a hurry that some peanuts wastes.", "reasoning_steps": [{"modality": "vision", "evidence": "At 8:05, the man got off the tractor for the first time.", "inference": "Sets the point to check the visuals."}, {"modality": "vision", "evidence": "At 8:10, the man was filling the bucket with peanuts.", "inference": "Peanuts are increasingly filling the bucket."}, {"modality": "vision", "evidence": "At 8:13, the bucket is full.", "inference": "The bucket cannot hold any more peanuts."}, {"modality": "audio", "evidence": "At 8:13, the man's voice increases suddenly.", "inference": "Sets the point to find out the reason why he's increasing his voice."}, {"modality": "audio", "evidence": "At 8:14, He said, 'Turn off! Turn off!'", "inference": "The man asks the machine operator to turn off the peanut transfer tube."}, {"modality": "vision", "evidence": "At 8:19, the peanuts keeps overfilling the bucket.", "inference": "Thus the reason why he raises his voice is that he worries about the waste of peanuts."}]}
{"id": "video_81__q0000", "video_id": "video_81", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_81_t5s.mp4"], "extra_info": {"question": "What is the tone of the person in black clothes when talking about drinking?"}, "reward_model_ground_truth": {"multi_choice": ["A. an excited tone", "B. a rhetorical tone", "A. an angry tone", "B. a cherished tone"], "correct_option": "A"}, "answer_text": "an excited tone", "reasoning_steps": [{"modality": "vision", "evidence": "find the man in black clothes"}, {"modality": "audio", "evidence": "At 02:17, the sous chef said 'the recipes for those things are more like chemistry equations.'", "inference": "they are talking about the past memory of drinking together."}, {"modality": "audio", "evidence": "At 02:51, the chef said 'gotcha, gotcha, and so you should be, it's delicious'.", "inference": "the chef was also excited about that memory."}]}
{"id": "video_81__q0001", "video_id": "video_81", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_81_t5s.mp4"], "extra_info": {"question": "Which dish did the actor who plays Iron Man clearly say he likes, made by the person who briefly showed the cooking scene?"}, "reward_model_ground_truth": {"multi_choice": ["A. Sticky pork bow bun.", "B. Sandwich.", "C. Chicken verinique", "D. Katsu curry."], "correct_option": "D"}, "answer_text": "Chicken verinique", "reasoning_steps": [{"modality": "vision", "evidence": "Around the video time of [8.05], a picture exists on the screen is chicken verinique.", "inference": "They are talking about chicken verinique."}, {"modality": "audio", "evidence": "Aftering showing the picture of chicken verinique,Tom says 'I love that'.", "inference": "The answer is chicken verinique."}]}
{"id": "video_81__q0002", "video_id": "video_81", "qid": 2, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_81_t5s.mp4"], "extra_info": {"question": "Which of the following is the correct order of the occurrence of these scenes?"}, "reward_model_ground_truth": {"multi_choice": ["A. Talk about athlete experience->Gordon introduces a canned seasoning->Tom stirs the shredded vegetables", "B. Talk about athlete experience->Tom stirs the shredded vegetables->Gordon introduces a canned seasoning", "A. Talk about dancing experience->Gordon introduces a canned seasoning->Tom stirs the shredded vegetables", "B. Talk about dancing experience->Tom stirs the shredded vegetables->Gordon introduces a canned seasoning"], "correct_option": "A"}, "answer_text": "Talk about athlete experience->Gordon introduces a canned seasoning->Tom stirs the shredded vegetables", "reasoning_steps": [{"modality": "audio", "evidence": "Before the video time [1.00], they have already started to talk about athlete experience.", "inference": "'Talk about athlete experience' is first."}, {"modality": "audio", "evidence": "Around the video time of [1.03], the host says 'but it's done with an incredible marinade'.", "inference": "The beer is an canned seasoning.'Gordon introduces a canned seasoning' is second."}, {"modality": "vision", "evidence": "Around the video time of [7.38], we can see Tom is stirring the vegetables.", "inference": "'Tom stirs the shredded vegetables' is third."}]}
{"id": "video_81__q0003", "video_id": "video_81", "qid": 3, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_81_t5s.mp4"], "extra_info": {"question": "What attitude does the the person wearing a red apron hold towards the reasons for his own first fame?"}, "reward_model_ground_truth": {"multi_choice": ["A. confident", "B. awkward", "A. curious", "B. surprised"], "correct_option": "D"}, "answer_text": "surprised", "reasoning_steps": [{"modality": "vision", "evidence": "find the person wearing a red apron"}, {"modality": "audio", "evidence": "At 03:02, the sous chef said 'what's really interesting is I never wanted to be an actor.'", "inference": "they are talking about the reasons for the sous chef's first fame."}, {"modality": "audio", "evidence": "At 03:41, the sous chef said 'headmaster of the school kind of told my parents you should get your son to audition for Billy Elliott.'", "inference": "they are talking about the sous chef's first fame."}, {"modality": "audio", "evidence": "At 03:48, the sous chef said 'it was completely by chance.'", "inference": "the sous chef was surprised at the reasons for his own first fame."}]}
{"id": "video_82__q0000", "video_id": "video_82", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_82_t5s.mp4"], "extra_info": {"question": "What color is the clothes worn inside the jacket of the person who organizes everyone in an orderly manner and urges them not to run when the fire alarm rings? "}, "reward_model_ground_truth": {"multi_choice": ["A.White", "B.Pink", "C.Green", "D.Black"], "correct_option": "C"}, "answer_text": "Green", "reasoning_steps": [{"modality": "audio", "evidence": "The fire alarm starts to ring at 10:45.", "inference": "To answer the question, we need to focus on this approximate time."}, {"modality": "audio", "evidence": "At 10:51, we can hear a female voice organizing everyone in an orderly manner and urges them not to run.", "inference": "This is the person mentioned in the question."}, {"modality": "vision", "evidence": "At 10:51, we can see a woman wearing a jacket.", "inference": "This is exactly the person mentioned in the question."}, {"modality": "vision", "evidence": "The color of her inner clothing is green.", "inference": "The color of the clothes worn inside the jacket of the woman who organizes everyone in an orderly manner when the fire alarm rings is green."}]}
{"id": "video_83__q0000", "video_id": "video_83", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_83_t5s.mp4"], "extra_info": {"question": "Which gang does the prisoner in front of the prison warden belong to when the warden mentions that prisoners usually have women's faces tattooed on their legs?"}, "reward_model_ground_truth": {"multi_choice": ["A.MS 13.", "B.Barrio 18.", "C.I don't know.", "D.MS 18."], "correct_option": "B"}, "answer_text": "Barrio 18.", "reasoning_steps": [{"modality": "audio", "evidence": "They mentioned at 19:07.", "inference": "Locate when they mention the prisoner."}, {"modality": "vision", "evidence": "There is an 18 logo on his back.", "inference": "Maybe the prisoner is Barrio 18."}, {"modality": "audio", "evidence": "Their gang can be inferred from their tattoos.", "inference": "So he is from Barrio 18."}]}
{"id": "video_83__q0001", "video_id": "video_83", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_83_t5s.mp4"], "extra_info": {"question": "Which gang does the  prisoner who once kidnapped the female athlete belong to?"}, "reward_model_ground_truth": {"multi_choice": ["A.MS 13.", "B.Barrio 18.", "C.I don't know.", "D.MS 18."], "correct_option": "C"}, "answer_text": "I don't know.", "reasoning_steps": [{"modality": "audio", "evidence": "They mentioned the prisoner at 17:36.", "inference": "Locate when they mention the prisoner."}, {"modality": "audio", "evidence": "They just mentioned that he is from a gang.", "inference": "The prisoner is from a gang."}, {"modality": "vision", "evidence": "There aren't any obvious logos on the prisoner's body.", "inference": "Can not find which one is the prisoner who once kidnapped the female athlete."}]}
{"id": "video_84__q0000", "video_id": "video_84", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_84_t5s.mp4"], "extra_info": {"question": "After the streamer said 'Bro, pick me up', who was the second person to die?"}, "reward_model_ground_truth": {"multi_choice": ["A.TheShy", "B.Elk.", "C.On.", "D.Meiko."], "correct_option": "A"}, "answer_text": "TheShy", "reasoning_steps": [{"modality": "audio", "evidence": "At 9:34, the streamer said ' Bro, pick me up'.", "inference": "Get the time point of the sentence."}, {"modality": "vision", "evidence": "At 9:53,  GALA killed On.", "inference": "The first to last person to die was On at 9:53."}, {"modality": "vision", "evidence": "Shadow killed TheShy at 11:15.", "inference": "The second to last person to die was TheShy at 11:15."}]}
{"id": "video_84__q0001", "video_id": "video_84", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_84_t5s.mp4"], "extra_info": {"question": "Before the streamer said 'Bin saved this team fight', who was the second to last person to die?"}, "reward_model_ground_truth": {"multi_choice": ["A.Wei.", "B.Elk.", "C.Rookie.", "D.Meiko."], "correct_option": "C"}, "answer_text": "Rookie.", "reasoning_steps": [{"modality": "audio", "evidence": "At 8:28, the streamer said 'Bin saved this team fight'.", "inference": "Get the time point of the team fight."}, {"modality": "vision", "evidence": "At 8:28, the screen shows that Triple Kill - Bin.", "inference": "Bin killed 3 enemies during the team fight."}, {"modality": "vision", "evidence": "At 8:26, Bin killed Wei, scoring a Triple Kill.", "inference": "The first to last person to die was Wei at 8:26."}, {"modality": "vision", "evidence": "At 8:23, Bin killed Rookie, scoring double kill.", "inference": "the second to last person to die was Rookie at 8:23."}]}
{"id": "video_85__q0000", "video_id": "video_85", "qid": 0, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_85_t5s.mp4"], "extra_info": {"question": "Among the following events, which one happened third in chronological order?"}, "reward_model_ground_truth": {"multi_choice": ["A.The blogger took a bite of waffle.", "B.The blogger took a bite of squid.", "C.The blogger ate a croissant.", "D.The blogger ate coconut rice."], "correct_option": "B"}, "answer_text": "The blogger took a bite of squid.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:15, the man said, 'I went to Paris where I had croissants.'", "inference": "he had croissants before."}, {"modality": "vision", "evidence": "At 1:46, the man took a bite of waffle.", "inference": "this was the second thing happened in chronological order."}, {"modality": "vision", "evidence": "At 4:03, the man took a bite of squid.", "inference": "so the third thing happened in chronological order is the man took a bite of squid."}]}
{"id": "video_86__q0000", "video_id": "video_86", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_86_t5s.mp4"], "extra_info": {"question": "What color is that old school Chevrolet?"}, "reward_model_ground_truth": {"multi_choice": ["A. Red.", "B. Yellow.", "C. Black.", "D. Grey."], "correct_option": "B"}, "answer_text": "Yellow.", "reasoning_steps": [{"modality": "audio", "evidence": "At 15:43, the man said, 'old school Chevrolet'.", "inference": "identify the old school Chevrolet."}, {"modality": "vision", "evidence": "At 15:43, in the background in front of the car window, the old school Chevrolet can be seen.", "inference": "identify the old school Chevrolet."}, {"modality": "vision", "evidence": "At 15:43, the old school Chevrolet in the picture is Yellow.", "inference": "identify the old school Chevrolet is Yellow."}]}
{"id": "video_86__q0001", "video_id": "video_86", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_86_t5s.mp4"], "extra_info": {"question": "What is the next landmark after the Manhattan Bridge?"}, "reward_model_ground_truth": {"multi_choice": ["A. The Empire State Building.", "B. The Broadway.", "C. The Canal Street.", "D. The Chinatown."], "correct_option": "A"}, "answer_text": "The Empire State Building.", "reasoning_steps": [{"modality": "audio", "evidence": "At 07:52, the man said, 'this is the Manhattan Bridge'.", "inference": "identify the Manhattan Bridge."}, {"modality": "vision", "evidence": "At 08:24, in the distant background outside the car window, Empire State Building could be seen.", "inference": "Empire State Building appeared."}, {"modality": "audio", "evidence": "At 08:24, the man said, 'Empire State Building.'", "inference": "identify the next landmark, Empire State Building."}]}
{"id": "video_87__q0000", "video_id": "video_87", "qid": 0, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_87_t5s.mp4"], "extra_info": {"question": "What happened to the team's coach during the game?"}, "reward_model_ground_truth": {"multi_choice": ["A. He was ill and unable to participate in the entire competition.", "B. He was sent off by the referee.", "C. He was provoked by the opposing players.", "D. He was collided with by a player."], "correct_option": "D"}, "answer_text": "He was collided with by a player.", "reasoning_steps": [{"modality": "vision", "evidence": "at 4:17, the man in black is a coach.", "inference": "get the person."}, {"modality": "audio", "evidence": "at 8:22, the man said ' hope his coach is well...'.", "inference": "get the person."}, {"modality": "vision", "evidence": "at 8:06, the coach is collided with by a player.", "inference": "He was collided with by a player."}]}
{"id": "video_88__q0000", "video_id": "video_88", "qid": 0, "question_type": "attribute comparison", "audio_type": "Music", "videos": ["data/videoomnibench/videos_downsampled/video_88_t5s.mp4"], "extra_info": {"question": "What is one of the characteristics of the second piece of music played in the car compared to the music when a blue car is in front of my car?"}, "reward_model_ground_truth": {"multi_choice": ["A. More relaxing.", "B. Top right.", "C. Louder volume.", "D. Higher pitch."], "correct_option": "C"}, "answer_text": "Louder volume.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:34, the first piece of music in the car starts.", "inference": "The first piece of music is played."}, {"modality": "audio", "evidence": "At 7: 34, the second piece of music starts.", "inference": "The second piece of music is played."}, {"modality": "audio", "evidence": "The volume of the second voice increased significantly", "inference": "Louder volume is one of the characteristics of the second piece of music played in the car compared to the first."}]}
{"id": "video_88__q0001", "video_id": "video_88", "qid": 1, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_88_t5s.mp4"], "extra_info": {"question": "What do I think has to come off while driving?"}, "reward_model_ground_truth": {"multi_choice": ["A. Slippers.", "B. Roof rack.", "C. Car windows.", "D. The mobile phone."], "correct_option": "A"}, "answer_text": "Slippers.", "reasoning_steps": [{"modality": "vision", "evidence": "At 3:17, the man is taking off his slippers.", "inference": "The man takes off his slippers before driving the car."}, {"modality": "audio", "evidence": "At 3: 19, the man's voice says'These slippers nee to come off first.'", "inference": "The man thinks slippers have to come off while driving."}]}
{"id": "video_89__q0000", "video_id": "video_89", "qid": 0, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_89_t5s.mp4"], "extra_info": {"question": "What was Daryl doing before I called him?"}, "reward_model_ground_truth": {"multi_choice": ["A. Attending the football team tryouts.", "B. Watching a football game.", "C. Attending an Amazon job interview.", "D. Recording a POV."], "correct_option": "A"}, "answer_text": "Attending the football team tryouts.", "reasoning_steps": [{"modality": "audio", "evidence": "'I' called a man Daryl at 0:30 in the video.", "inference": "This is the 'Daryl' in the question."}, {"modality": "audio", "evidence": "'I' said Daryl was back from football at 0:33.", "inference": "We know Daryl was back from football."}, {"modality": "vision", "evidence": "The text 'left for 2 weeks to play amateur football' was displayed below the video screen at 0:36.", "inference": "We know Daryl was back from playing amateur football."}, {"modality": "audio", "evidence": "'I' said he didn't make the cut at 0:38.", "inference": "We know Daryl was attending the football team tryouts, but he didn't make the cut. So he was back."}]}
{"id": "video_89__q0001", "video_id": "video_89", "qid": 1, "question_type": "ego reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_89_t5s.mp4"], "extra_info": {"question": "After exiting lane five, where should I turn?"}, "reward_model_ground_truth": {"multi_choice": ["A. The left.", "B. The right.", "C. The front.", "D. Turning around."], "correct_option": "B"}, "answer_text": "The right.", "reasoning_steps": [{"modality": "audio", "evidence": "'I' said lane 5 goes first at 7:10.", "inference": "We know 'I' was going to drive towards lane 5 at that moment."}, {"modality": "audio", "evidence": "'I' said, 'After lane 5, we are out' at 7:14.", "inference": "We know 'I' would exit lane 5."}, {"modality": "vision", "evidence": "As 'I' said, we can see 'I' was driving towards lane 5 at 7:19.", "inference": "We see 'I' was driving towards lane 5."}, {"modality": "vision", "evidence": "As 'I' said, we can see 'I' was exiting lane 5 at 7:25.", "inference": "We see 'I' was exiting lane 5."}, {"modality": "vision", "evidence": "At 7:27, I turned the steering wheel to the right and drove to the right..", "inference": "After exiting lane five, I turned to the right."}]}
{"id": "video_89__q0002", "video_id": "video_89", "qid": 2, "question_type": "summarization", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_89_t5s.mp4"], "extra_info": {"question": "What tips did Khan help demonstrate?"}, "reward_model_ground_truth": {"multi_choice": ["A. Saveing space inside the car.", "B. Easily unloading cargo from the car.", "C. Easily loading cargo into the car.", "D. Opening the compartment."], "correct_option": "D"}, "answer_text": "Opening the compartment.", "reasoning_steps": [{"modality": "audio", "evidence": "'I' said we got my boy Khan at 1:58.", "inference": "This is the 'Khan' in the question."}, {"modality": "vision", "evidence": "We can see the face of Khan at 1:58.", "inference": "We know what Khan looks like."}, {"modality": "audio", "evidence": "'I' mentioned one of the tips 'I' was gonna show today at 2:25.", "inference": "This might be the tip mentioned in the question."}, {"modality": "vision", "evidence": "We can see Khan opened the compartment at 2:35.", "inference": "So we know Khan helped demonstrate the tip of opening the compartment."}]}
{"id": "video_9__q0000", "video_id": "video_9", "qid": 0, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_9_t5s.mp4"], "extra_info": {"question": "What emotion did the person in pink have on the short expert at first?"}, "reward_model_ground_truth": {"multi_choice": ["A. Interested", "B. Happy", "C. Indifferent", "D. Dislike"], "correct_option": "A"}, "answer_text": "Interested", "reasoning_steps": [{"modality": "vision", "evidence": "The left woman wears pink clothes.", "inference": "get the person in pink."}, {"modality": "vision", "evidence": "at 1:08,the short expert is on the right.", "inference": "get the person."}, {"modality": "audio", "evidence": "At 1:09,the short expert  said to the woman in pink that she is insecure.", "inference": "get the words he said."}, {"modality": "audio", "evidence": "At 1:14,the person in pink said 'omg, she is loving that.'", "inference": "she is interested."}]}
{"id": "video_9__q0001", "video_id": "video_9", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_9_t5s.mp4"], "extra_info": {"question": "What emotion did the black person have on the short expert after he showed his ability?"}, "reward_model_ground_truth": {"multi_choice": ["A. Interested", "B. Dislike", "C. Indifferent", "D. Doubtful"], "correct_option": "D"}, "answer_text": "Doubtful", "reasoning_steps": [{"modality": "vision", "evidence": "The woman on the right is a black person.", "inference": "get the person."}, {"modality": "vision", "evidence": "at 1:08,the short expert is on the right.", "inference": "get the person."}, {"modality": "audio", "evidence": "At 2:01,the short expert shows his ability.", "inference": "get the words he said."}, {"modality": "audio", "evidence": "At 2:06,the black person said 'sorry are you sure that you two are professional？'", "inference": "she is doubtful."}]}
{"id": "video_90__q0000", "video_id": "video_90", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_90_t5s.mp4"], "extra_info": {"question": "How are the orange tools made?"}, "reward_model_ground_truth": {"multi_choice": ["A. Made with 3D printing.", "B. Purchased from the official repair factory.", "C. Made in my own factory.", "D. Obtained by repairing old tools."], "correct_option": "A"}, "answer_text": "Made with 3D printing.", "reasoning_steps": [{"modality": "audio", "evidence": "At 04:09, the man said, 'quick little ratchet flip tutorial for you, bud'.", "inference": "he was going to do the action of quickly flipping the small ratchet."}, {"modality": "audio", "evidence": "At 04:13, the man said, 'you go like this.'", "inference": "he was doing one part of the action."}, {"modality": "vision", "evidence": "At 04:13, the man was flipping the ratchet forward 1 time.", "inference": "the first part of the action is 1 forward flip."}, {"modality": "audio", "evidence": "At 04:14, the man said, 'then after you do that you go like that.'", "inference": "he was doing the other part of the action."}, {"modality": "vision", "evidence": "At 04:14, the man was flipping the ratchet backward 1 time.", "inference": "the second part of the action is 1 backward flip. So the action consists of 1 forward flip and 1 backward flip."}]}
{"id": "video_92__q0000", "video_id": "video_92", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_92_t5s.mp4"], "extra_info": {"question": "Which two people were performing the action for the word that the group incorrectly guessed as 'Tiger Circle'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Chen He and Gong Jun.", "B. Ouyang Didi and Liu Yuning.", "C. Sun Honglei and Li Naiwen.", "D. Gong Jun and Wang Yuwen."], "correct_option": "D"}, "answer_text": "Gong Jun and Wang Yuwen.", "reasoning_steps": [{"modality": "audio", "evidence": "At 4:16, 'Tiger Circle' is mentioned by a man.", "inference": "To check if this word is incorrect."}, {"modality": "vision", "evidence": "At the same momemt, the upper left corner of the screen shows the correct answer: Coloseuum.", "inference": "So the man guessed it incorrectly."}, {"modality": "audio", "evidence": "At 3:55, the subtitle shows the names of the performers: Gong Jun and Wang Yuwen.", "inference": "Gong Jun and Wang Yuwen were performing the action for the word that the group incorrectly guessed as 'Tiger Circle'"}]}
{"id": "video_92__q0001", "video_id": "video_92", "qid": 1, "question_type": "temporal understanding", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_92_t5s.mp4"], "extra_info": {"question": "In the first round of the warm-up morning exercise game, who was eliminated by the command 'Seafood Lemon'?"}, "reward_model_ground_truth": {"multi_choice": ["A. Ouyang Didi.", "B. Wang Yuwen.", "C. Li Naiwen.", "D. Gong Jun."], "correct_option": "C"}, "answer_text": "Li Naiwen.", "reasoning_steps": [{"modality": "audio", "evidence": "At 9:15,  the narrator says: We use 3 kinds of local food, including lemon, seafood, pizza, to design a morning exercise.", "inference": "To check when the command 'Seafood Lemon' happens."}, {"modality": "audio", "evidence": "At 18:37, the command 'Seafood Lemon' is mentioned.", "inference": "To check who's is eliminated after this."}, {"modality": "vision", "evidence": "At 18:42, Li Naiwen did the wrong action.", "inference": "So Li Naiweng is eliminated after the command 'Seafood Lemon' ."}]}
{"id": "video_93__q0000", "video_id": "video_93", "qid": 0, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_93_t5s.mp4"], "extra_info": {"question": "Which of these lines was spoken by the person who gets carsick?"}, "reward_model_ground_truth": {"multi_choice": ["A. Perhaps among the children who play football, there are world superstars.", "B. This is really a bag of money.", "C. Sometimes I just have to step forward.", "D. If the money is 26,000, go back directly."], "correct_option": "D"}, "answer_text": "If the money is 26,000, go back directly.", "reasoning_steps": [{"modality": "vision", "evidence": "the person who says they are carsick."}, {"modality": "audio", "evidence": "listen to the dialogue spoken by this person."}, {"modality": "audio", "evidence": "match one of their lines to the provided options."}]}
{"id": "video_93__q0001", "video_id": "video_93", "qid": 1, "question_type": "fine-grained perception", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_93_t5s.mp4"], "extra_info": {"question": "What is the nickname of the person who says they are afraid of phone calls in group chats?"}, "reward_model_ground_truth": {"multi_choice": ["A. Na Xiaoda.", "B. Chen Xia 'er.", "C. Li Laowu.", "D. Zhang Laoqi."], "correct_option": "D"}, "answer_text": "Zhang Laoqi.", "reasoning_steps": [{"modality": "audio", "evidence": "identify the person who says they are afraid of phone calls in group chats."}, {"modality": "vision", "evidence": "locate the segment where nicknames are being discussed."}, {"modality": "audio", "evidence": "identify the nickname given to that person."}]}
{"id": "video_94__q0000", "video_id": "video_94", "qid": 0, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_94_t5s.mp4"], "extra_info": {"question": "The person who said 'down' as the enemy members were leaving, which group did they depart with?"}, "reward_model_ground_truth": {"multi_choice": ["A. First.", "B. Second.", "C. Third.", "D. Fourth."], "correct_option": "D"}, "answer_text": "Fourth.", "reasoning_steps": [{"modality": "audio", "evidence": "At 6:36, Zheng Kai said 'down'.", "inference": "Pay attention to him that which group did he depart with."}, {"modality": "vision", "evidence": "At 6:15, Zheng Kai was running towards the center of the pool.", "inference": "Pay attention to which group displayed in the scene."}, {"modality": "vision", "evidence": "At 6:15, the text appeared in the upper-middle part of the scene: 'The fourth group, set off.'", "inference": "So the answer is 'Fourth.'."}]}
{"id": "video_94__q0001", "video_id": "video_94", "qid": 1, "question_type": "reference reasoning", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_94_t5s.mp4"], "extra_info": {"question": "Who was the first person to give up looking for something because of the heat?"}, "reward_model_ground_truth": {"multi_choice": ["A. Sha Yi.", "B. Chen Zheyuan.", "C. Li Yunrui.", "D. Bai Lu."], "correct_option": "D"}, "answer_text": "Bai Lu.", "reasoning_steps": [{"modality": "audio", "evidence": "At 2:10, a woman said 'It's so hot...'.", "inference": "This indicates that this woman might give up looking for the item."}, {"modality": "vision", "evidence": "At 2:10, Bai Lu was show in the scene.", "inference": "She might be the first person to give up looking for something because of heat."}, {"modality": "vision", "evidence": "At 2:12, there was text 'give up looking for' shown beside Bai Lu.", "inference": "So the answer is 'Bai Lu'."}]}
{"id": "video_94__q0002", "video_id": "video_94", "qid": 2, "question_type": "spatial understanding", "audio_type": "Sound", "videos": ["data/videoomnibench/videos_downsampled/video_94_t5s.mp4"], "extra_info": {"question": "Which seat was occupied by the person sitting behind the one who had applied sunscreen, relative to the person who was reviewing the film?"}, "reward_model_ground_truth": {"multi_choice": ["A. The third person on the right hand side.", "B. The second person on the right hand side.", "C. The first one on the right hand side.", "D. The first one on the left hand side."], "correct_option": "C"}, "answer_text": "The third person on the right hand side.", "reasoning_steps": [{"modality": "vision", "evidence": "At 4:42, a woman with a high ponytail in distance was applying sunscreen.", "inference": "Identify her."}, {"modality": "audio", "evidence": "At 4:23, a man asked the person 'Are you reviewing the film?,and he is the chief director", "inference": "Identify him."}, {"modality": "vision", "evidence": "At 4:42, the woman was on the chief director's right hand side.", "inference": "Pay attention to which person she was (from the left) on the right hand side of the chief director."}, {"modality": "vision", "evidence": "At 4:42, the woman was the third person on the chief director's right hand side.", "inference": "So the answer is 'The third person on the right hand side.'."}]}
{"id": "video_95__q0000", "video_id": "video_95", "qid": 0, "question_type": "counting", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_95_t5s.mp4"], "extra_info": {"question": "In the teasing game with five people in one round, how many of them are genuinely giving compliments?"}, "reward_model_ground_truth": {"multi_choice": ["A. 0.", "B. 1.", "C. 2.", "D. 3."], "correct_option": "C"}, "answer_text": "2.", "reasoning_steps": [{"modality": "vision", "evidence": "At 0:32, the text 'Five-person game' is shown in the center of the screen.", "inference": "This might be the teasing game with five people in one round mentioned in the question."}, {"modality": "audio", "evidence": "At 0:35, we can hear a male voice saying, 'Please share your teases for the show's production team.'", "inference": "This should be the teasing game with five people in one round mentioned in the question."}, {"modality": "audio", "evidence": "At 0:52, we can hear a male voice saying, 'Yao Yitian has too many fans.'", "inference": "A man offered a friendly 'tease'."}, {"modality": "audio", "evidence": "At 0:56, another male voice doubted, 'Is it a tease?'", "inference": "Based on this response, it can be inferred that the 'tease' is actually a compliment."}, {"modality": "audio", "evidence": "At 1:09, we can hear a new male voice saying, 'Yao PD is too handsome.'", "inference": "Here is the other man offering a friendly 'tease'."}, {"modality": "vision", "evidence": "At 1:15, the people around couldn't help laughing.", "inference": "Based on other people's responses, it can be inferred that this 'tease' is also a compliment. In conclusion, there are 2 people genuinely giving compliments in the teasing game."}]}
{"id": "video_96__q0000", "video_id": "video_96", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_96_t5s.mp4"], "extra_info": {"question": "What did the tallest person here say?"}, "reward_model_ground_truth": {"multi_choice": ["A. What you said should not imply any connection with socks.", "B. I once thought that the most dangerous place was the safest place.", "C. I suspect that some people have already completed some actions.", "D. From now on, no one should touch me."], "correct_option": "D"}, "answer_text": "From now on, no one should touch me.", "reasoning_steps": [{"modality": "audio", "evidence": "At 09:21, Liu Yuning held up his mobile phone and said, 'It's at a height like a surveillance camera.'", "inference": "identify the tallest person is Liu Yuning."}, {"modality": "vision", "evidence": "At 09:21, Liu Yuning is the tallest person in the picture.", "inference": "identify the tallest person is Liu Yuning."}, {"modality": "audio", "evidence": "At 08:02, Liu Yuning said, 'From now on, no one should touch me.'", "inference": "identify what the tallest person said was 'From now on, no one should touch me.'"}]}
{"id": "video_97__q0000", "video_id": "video_97", "qid": 0, "question_type": "attribute comparison", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_97_t5s.mp4"], "extra_info": {"question": "How did the exclusive butter-cooking method compare to expectations for a well-cooked lobster?"}, "reward_model_ground_truth": {"multi_choice": ["A. The method successfully enhanced the lobster's flavor and texture, meeting or exceeding expectations.", "B. The method yielded a cooked lobster, but the high butter content negatively impacted its texture, making it tough.", "C. The method resulted in a cooked lobster, but the extensive use of butter did not deliver a noticeable improvement in its quality or taste.", "D. The method completely failed to cook the lobster, leaving it raw and inedible."], "correct_option": "C"}, "answer_text": "The method resulted in a cooked lobster, but the extensive use of butter did not deliver a noticeable improvement in its quality or taste.", "reasoning_steps": [{"modality": "audio", "evidence": "At 0:13, a man said 'The real question is, how does it taste?'.", "inference": "The lobster cooked in butter is about to be tasted."}, {"modality": "vision", "evidence": "In the scene at 0:13, a man was tasting the claw of the lobster.", "inference": "It is very likely that he would give a comment after tasting it."}, {"modality": "audio", "evidence": "At 0:14, the man said 'That did absolutely nothing.'.", "inference": "This indicates that lobster cooked in butter tastes no different from lobster prepared in other ways."}]}
{"id": "video_97__q0001", "video_id": "video_97", "qid": 1, "question_type": "sentiment analysis", "audio_type": "Speech", "videos": ["data/videoomnibench/videos_downsampled/video_97_t5s.mp4"], "extra_info": {"question": "What was the mood of the boy on the left before melting the butter?"}, "reward_model_ground_truth": {"multi_choice": ["A.happy", "B.sad", "C.sorry", "D.excited"], "correct_option": "C"}, "answer_text": "sorry", "reasoning_steps": [{"modality": "vision", "evidence": "The boys on the right begins talk at the beginning.", "inference": "The voice at the beginning belongs to the  boy on the right."}, {"modality": "audio", "evidence": "The voice says ' let's cook the lobster using only butter'", "inference": "identify the voice of the boy on the right."}, {"modality": "audio", "evidence": "Then another voice says 'I'm sorry' at 0:03.", "inference": "This voice comes from the boy on the left."}, {"modality": "vision", "evidence": "Then butter is melting at 0:05", "inference": "The mood of the boy on the left before melting the butter is sorry."}]}