I work with multitrack sequences, and I know about all of the work-arounds like removing cross-talk and making separate compositions, but Descript still regularly attributes the wrong speaker to my corrections, even if it is on an entirely different track for which I specifically said was single-speaker.

It would be so much more efficient for me if I could just highlight a portion of a track in the multi-track view, type in what the audio says, and have Descript me rather than try to be "smart" and attempt to align the text with nearby words or whatever.

Thanks!