On-the-Fly Conversion of FFmpeg-Supported Audio Files

[SilverServer34]

Since the software already uses ffmpeg internally, it should be fairly straightforward to convert compatible audio files on the fly for transcription.

For example, to convert an audio file extracted from WhatsApp (which is typically saved in .opus format), you can simply run:

ffmpeg -i "input.opus" -vn -c:a libmp3lame -b:a 192k -ar 48000 -map_metadata 0 -id3v2_version 3 "${input%.opus}.mp3"

I created a Quick Action on macOS that does exactly this — it converts the selected .opus file to .mp3 using ffmpeg. Then I just drag the resulting MP3 into WhisperDesk for transcription.

[pedrovsiqueira]

Ok I will take a look and implement on version 1.9.0. I will keep you updated.

[pedrovsiqueira]

@SilverServer34 I have deployed to version 1.9.0. .opus format will now work as well.

[SilverServer34]

That's awesome, thank you! Just tried it and it works beautifully.