how to select Audio Track ?

[fullyconv]

I have a video with multiple Audio Track (like french english ...etc) How can i transcribe a spesific language from all of these languages?

[abdeladim-s]

Unfortunately the whisper model has no built-in method to do this currently. What you could do though is to extract the video with each audio track separately into multiple files using ffmpeg, then run the CLI to batch process them all, or run them manually one by one using the webui.

You can extract the video and the first audio track for example using this command:

ffmpeg -i input-video.mp4 -map 0:v:0 -map 0:a:0 out-en.mp4