I have an interview clip with two speakers. When transcribing, Premiere correctly identifies the interviewer as a single speaker (Speaker 3) but assigns the interviewee two speakers (Speaker 1 and Speaker 2), even though it's clearly the same voice throughout.
Moreover, it constantly breaks up the interviewee's sentences between 1 and 2, even after just a few words.
Here's a screen cap of what's happening: https://youtu.be/OtJxeWa0bhg. It's playing from the source window, so it's including the on-camera mics.
This fragmentation makes it harder to review and edit within the text window. Even when I manually reassign both sides of a split to the same speaker, the break carries over when generating captions. So I end up with a bunch of unnecessary caption divisions.
I know Premiere sometimes fails to distinguish between speakers, but this is the first time I've seen it over-assign speakers to one person and rapidly switch between them. Has anyone else encountered this, and is there a fix or workaround?
The clip in question is a 1.5 hour .MXF file from an FX6 with four audio channels. Channel 1 is a boom and 2-4 are in-camera. I transcribe using Channel 1. I'm running Premiere 25.2.3 on an Apple M3 MAX MacBook Pro using Sonoma 14.7.5.
AdChoices