Sometimes when I use the detect speakers function (in creating and editing a transcript), the program automatically pulls a section of audio with a lot of crosstalk/people speaking over one another. This makes it difficult to tell which voice the program has identified for a section of audio, and the speaker labels get attributed to the wrong person.

PLEASE ensure that when the program is pulling audio for speaker detection, it is pulling an isolated section of audio with one voice at a time, as optimal audio quality as possible, and clearly referring to one speaker.