Add listening demo section (background noise, emotion, speaking style, accent)
Browse files
README.md
CHANGED
|
@@ -38,6 +38,37 @@ pipeline_tag: text-to-speech
|
|
| 38 |
|
| 39 |
Raon-OpenTTS is an open-data, open-weight zero-shot TTS system that performs on par with state-of-the-art closed-data models. This is the 1B variant.
|
| 40 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
## Key Features
|
| 42 |
|
| 43 |
- **Fully Open**: Both model weights and training data (615K hours, 11 English speech datasets) are publicly available for reproducible TTS research.
|
|
|
|
| 38 |
|
| 39 |
Raon-OpenTTS is an open-data, open-weight zero-shot TTS system that performs on par with state-of-the-art closed-data models. This is the 1B variant.
|
| 40 |
|
| 41 |
+
## Listening Demo
|
| 42 |
+
|
| 43 |
+
Zero-shot voice cloning samples from Raon-OpenTTS-1B across challenging conditions. Each row shows the generated audio, the input text, and the reference (prompt) audio.
|
| 44 |
+
|
| 45 |
+
### Background Noise
|
| 46 |
+
|
| 47 |
+
| Generated | Text | Reference |
|
| 48 |
+
|---|---|---|
|
| 49 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/EMOTIONAL/crema-d/crema-d_2261.wav"></audio> | That is exactly what happened | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/EMOTIONAL/crema-d/crema-d_4058.wav"></audio> |
|
| 50 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/NOISY/tedlium3_test/tedlium3_test_201.wav"></audio> | relative to how much i think of other vacations and | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/NOISY/tedlium3_test/tedlium3_test_800.wav"></audio> |
|
| 51 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/NOISY/librispeech_other/7902-96595-0006.wav"></audio> | I hope not, Dick; I hope not; but smugglers don't stand at anything sometimes". | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/NOISY/librispeech_other/7902-96595-0018.wav"></audio> |
|
| 52 |
+
|
| 53 |
+
### Emotion
|
| 54 |
+
|
| 55 |
+
| Generated | Text | Reference |
|
| 56 |
+
|---|---|---|
|
| 57 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/EMOTIONAL/EmoV_DB/EmoV_DB_522.wav"></audio> | To my surprise he began to show actual enthusiasm in my favor. | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/EMOTIONAL/EmoV_DB/EmoV_DB_337.wav"></audio> |
|
| 58 |
+
|
| 59 |
+
### Speaking Style
|
| 60 |
+
|
| 61 |
+
| Generated | Text | Reference |
|
| 62 |
+
|---|---|---|
|
| 63 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/EMOTIONAL/EmoV_DB/EmoV_DB_964.wav"></audio> | Keep an eye on him. | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/EMOTIONAL/EmoV_DB/EmoV_DB_797.wav"></audio> |
|
| 64 |
+
|
| 65 |
+
### Accent
|
| 66 |
+
|
| 67 |
+
| Generated | Text | Reference |
|
| 68 |
+
|---|---|---|
|
| 69 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/CLEAN/CMU_Arctic/CMU_Arctic_373.wav"></audio> | MacDougall tapped his forehead suspiciously with a stubby forefinger. | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/CLEAN/CMU_Arctic/CMU_Arctic_619.wav"></audio> |
|
| 70 |
+
| <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/gen_result_0408/1b/CLEAN/L2Arctic/L2Arctic_166.wav"></audio> | it occurred to me that there would have to be an accounting | <audio controls src="https://datacomptts-evaluation.s3.ap-northeast-2.amazonaws.com/eval_data_0408/ref_audio/CLEAN/L2Arctic/L2Arctic_174.wav"></audio> |
|
| 71 |
+
|
| 72 |
## Key Features
|
| 73 |
|
| 74 |
- **Fully Open**: Both model weights and training data (615K hours, 11 English speech datasets) are publicly available for reproducible TTS research.
|