mirror of
https://github.com/open-webui/open-webui.git
synced 2026-05-06 10:58:17 -05:00
[GH-ISSUE #14066] issue: The text obtained using the microphone is only in Traditional Chinese #32657
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @weiding9527 on GitHub (May 20, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/14066
Check Existing Issues
Installation Method
Docker
Open WebUI Version
0.6.10
Ollama Version (if applicable)
0.6.6
Operating System
centos7.9
Browser (if applicable)
quark夸克
Confirmation
README.md.Expected Behavior
I selected Simplified Chinese as the language in the open webUI system settings. Then, I used the microphone in the dialog box to capture my voice. After clicking ✅, it should convert the speech into Simplified Chinese.
我在open webUI系统设置选择语言为简体中文。然后在对话框中使用麦克风获取我的声音,点击✅之后,它应该把语音转换为简体中文
Actual Behavior
I selected Simplified Chinese as the language in the Open WebUI system settings. Then, I used the microphone in the dialog box to capture my voice, and after clicking ✅, it converted the speech into Traditional Chinese. But what I want is Simplified Chinese.
我在open webUI系统设置选择语言为简体中文。然后在对话框中使用麦克风获取我的声音,点击✅之后,它把语音转换成了繁体中文。但是我想要的是简体中文
Steps to Reproduce
Click "New Conversation", click the microphone icon, say a sentence, and click ☑️.
点击“新对话”,点击麦克风图标,说出一段话,点击☑️
Logs & Screenshots
2025-05-20 09:32:25.804 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:46762 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:35.315 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:34044 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:41.519 | INFO | open_webui.routers.audio:transcription:902 - file.content_type: audio/webm;codecs=opus - {}
2025-05-20 09:32:41.519 | INFO | open_webui.routers.audio:transcribe:781 - transcribe: /app/backend/data/cache/audio/transcriptions/ffe09b2e-12b8-4189-b24e-9d588d6d2108.webm - {}
Chunk paths: ['/app/backend/data/cache/audio/transcriptions/ffe09b2e-12b8-4189-b24e-9d588d6d2108.webm']
2025-05-20 09:32:41.933 | INFO | faster_whisper.transcribe:transcribe:839 - Processing audio with duration 00:07.200 - {}
2025-05-20 09:32:42.429 | INFO | faster_whisper.transcribe:transcribe:906 - Detected language 'zh' with probability 0.98 - {}
2025-05-20 09:32:42.430 | INFO | open_webui.routers.audio:transcription_handler:548 - Detected language 'zh' with probability 0.981205 - {}
2025-05-20 09:32:42.935 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:34164 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:42.959 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:51002 - "POST /api/v1/audio/transcriptions HTTP/1.0" 200 - {}
2025-05-20 09:32:46.347 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:33272 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:47.827 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.2.193:58199 - "GET /_app/version.json HTTP/1.1" 200 - {}
2025-05-20 09:32:47.863 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:44446 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:50.750 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:46134 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:50.847 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:55130 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:51.078 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:35464 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:52.656 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:39510 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:32:59.374 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:39032 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:02.498 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:42232 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:07.580 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:47480 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:09.865 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:51108 - "GET /_app/version.json HTTP/1.0" 200 - {}
2025-05-20 09:33:13.226 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:38636 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:16.068 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:51986 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:18.982 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:43136 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:27.459 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:55474 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:29.681 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:46572 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:31.950 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:49476 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:39.518 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:40344 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:48.836 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.2.193:58431 - "GET /_app/version.json HTTP/1.1" 200 - {}
2025-05-20 09:33:50.936 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:46182 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
2025-05-20 09:33:52.365 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 10.232.1.2:54710 - "GET /webman/resources/images/2x/default_login_background/dsm7_01.jpg HTTP/1.1" 200 - {}
Additional Information
No response
@ExDragine commented on GitHub (May 20, 2025):
The Whisper model, as an multilingual ASR model, first determines the language of what you are saying before recognising your speech content, and since Traditional Chinese is highly similar to Simplified Chinese, it is possible that the whisper model itself may confuse the two languages. The solution to this problem is to pass an expected language value before the actual transcription, which is set to 'auto' by default to be inferred by whisper itself, or it can be set by the user or kept in sync with the OpenWebUI language. A more radical solution would be to have an option to set the language by the user before transcription.
Overall, this is not a problem with OpenWebUI, but we can improve it.
@tjbck commented on GitHub (May 23, 2025):
Model issue, related: #13989