search

thewh1teagle / vibe

Transcribe on your own!
https://thewh1teagle.github.io/vibe/
MIT License
1.28k stars 73 forks source link

[Feature Request]: Transcribe only segment(s) of an audio file #329

Open oleole39 opened 1 month ago

oleole39 commented 1 month ago

Describe the feature

Hello,

When submitting an audio file for transcription, it will currently get transcribed from the beginning to the end (except if cancelled at some point).

If you need only part of the file to be transcribed, you can of course cut the file to the required segment priori to submitting it to Vibe, but it could prove convenient to be able define a segment from within Vibe, e.g. enter start time and end time as in ffmpeg -ss 00:22:10 -i file.mp3 -to 00:26:20 -c:a copy file_segment.mp3 so that it then starts transcription only for file_segment.mp3.

An even nicer but more complex implementation could also show the spectrogram of the submitted audio file and allow to select on it one or several portions to be transcribed (as in Audacity for instance).

thewh1teagle commented 1 month ago

Hey @oleole39 Thanks for creating the feature request : - )

I created a draft PR, but I'm unsure how useful the time range feature is. I'm only adding what's essential to keep the app minimal. https://github.com/thewh1teagle/vibe/pull/330

oleole39 commented 1 month ago

I created a draft PR, but I'm unsure how useful the time range feature is. I'm only adding what's essential to keep the app minimal. #330

Yes I understand. That draft PR looks good already!

thewh1teagle commented 2 weeks ago

You may be able to use the new custom ffmpeg command for that added in https://github.com/thewh1teagle/vibe/releases/tag/v2.6.7

oleole39 commented 1 week ago

Thanks, not as convenient for the current use case as what was discussed in draft design PR #330, but it is a way to it (and there is no absolute/urgent need for it).