#7781 No audio file was generated during the listening test.self=OmniVoice(uuid=None, proxy_str=None, last_down_time=0, tts_ty

42.1**0 Posted at: 8 hours ago

No audio file was generated during the listening test.
self=OmniVoice(uuid=None, proxy_str=None, last_down_time=0, tts_type=3, len=1, language='de', play=True, volume='+33%', rate='+0%', pitch='+0Hz', has_done=0, wait_sec=1.0, dub_nums=1, error='No audio file was generated during the listening test.', api_url='', model_name='', is_cuda=False, local_dir=None, is_redubb=False):
{'text': 'Hallo, mein lieber Freund, ich hoffe, dass jeder deiner Tage wunderbar und erfreulich ist!', 'rate': '+0%', 'role': 'zh_female_nverguo.wav', 'filename': 'D:/win-pyvideotrans-v4.14/tmp/16096/listen_voice/3-de-1790951894.6836991.wav', 'tts_type': 3, 'language': 'de', 'volume': '+33%', 'pitch': '+0Hz'}
Traceback (most recent call last):
File "videotrans\util\ListenVoice.py", line 21, in run
File "videotrans\tts\__init__.py", line 119, in run
File "videotrans\tts\_base.py", line 123, in run
videotrans.configure.excepts.DubbingSrtError: No audio file was generated during the listening test.
self=OmniVoice(uuid=None, proxy_str=None, last_down_time=0, tts_type=3, len=1, language='de', play=True, volume='+33%', rate='+0%', pitch='+0Hz', has_done=0, wait_sec=1.0, dub_nums=1, error='No audio file was generated during the listening test.', api_url='', model_name='', is_cuda=False, local_dir=None, is_redubb=False)

=
system:Windows-10-10.0.26200-SP0
version:v4.14
frozen:True
language:en_US
root_dir:D:/win-pyvideotrans-v4.14

#1 DeepSeek reply at 8 hours ago

The OmniVoice TTS channel failed to produce any output audio file during the listening test — meaning the model either did not load correctly or the voice-cloning reference was rejected, so no wav was written to tmp/.../listen_voice/.

Most likely causes and fixes (check in this order):

Model not fully downloaded / missing files
OmniVoice is a built-in channel; its model is not bundled and is downloaded on first use. Partial or failed downloads are the most common cause of "no audio file was generated".
- Verify the model exists at: D:/win-pyvideotrans-v4.14/models/models--k2-fsa--OmniVoice
- If missing or incomplete, delete the folder and manually download from https://huggingface.co/k2-fsa/OmniVoice/tree/main into that exact path.
- Reference: /aboutmodels
Reference audio problem (you selected zh_female_nverguo.wav)
Voice cloning requires a clean 3–10 second single-speaker WAV with no background noise and no leading/trailing silence.
- Put the file in the f5-tts folder of the software directory.
- Register it via Menu -> TTS Settings -> Set Reference Audio in the format filename.wav#the spoken text in the audio.
- Then re-select it in the voice role dropdown.
Blocked or unstable network / proxy
The model download goes through huggingface.co. Ensure your proxy is valid and working; an invalid proxy string causes all requests to fail. Clear the proxy box if you don't need one.
Insufficient memory / GPU conflict
is_cuda=False — if you have only 8GB RAM or less, try enabling CUDA (requires NVIDIA + CUDA 12.8 + cuDNN 9.11) or use a lighter channel such as Edge-TTS / ZipVoice / Qwen3-TTS to confirm the pipeline works.

Quick verification:

Test the same text with Edge-TTS(免费) — if that produces audio, the issue is isolated to the OmniVoice model/reference audio.
Check D:/win-pyvideotrans-v4.14/logs/ for the latest log file and read the last ~30 lines for the underlying model-loading error.

If the above does not resolve it, please provide the latest log file content (last ~30 lines) from the logs/ folder.

#7781 No audio file was generated during the listening test.self=OmniVoice(uuid=None, proxy_str=None, last_down_time=0, tts_ty

1 Replies

请查阅相关文档：

Post Your Reply

Related resource