Update modules/vad/silero_vad.py

modules/vad/silero_vad.py

@@ -252,22 +252,7 @@ class SileroVAD:
         ts_map = SpeechTimestampsMap(speech_chunks, sampling_rate)
 
         for segment in segments:
-            if segment["words"]:
-                words = []
-                for word in segment["words"]:
-                    # Ensure the word start and end times are resolved to the same chunk.
-                    middle = (word["start"] + word["end"]) / 2
-                    chunk_index = ts_map.get_chunk_index(middle)
-                    word["start"] = ts_map.get_original_time(word["start"], chunk_index)
-                    word["end"] = ts_map.get_original_time(word["end"], chunk_index)
-                    words.append(word)
-
-                segment["start"] = words[0].start
-                segment["end"] = words[-1].end
-                segment["words"] = words
-
-            else:
-                segment["start"] = ts_map.get_original_time(segment["start"])
-                segment["end"] = ts_map.get_original_time(segment["end"])
+            segment["start"] = ts_map.get_original_time(segment["start"])
+            segment["end"] = ts_map.get_original_time(segment["end"])
 
         return segments