Multichannel WAV for transcribing podcasts with multiple speakers

[rstockm]

Recording each speaker of a podcast on a separate track in a DAW has lot's of benefits for the post-production (reduce crosstalk, tune voices mit eq etc.). It's also quite easy to export a multichannel WAV file from Audition, REAPER etc. - which would be a perfect input for Whisper. Each channel contains one voice, but within only 1 file to handle.

Workflow:

• import multichannel WAV
• separate the tracks
• transcribe each track separately with Whisper
• merge the transcriptions via timecode and add speaker information

done!