Would be very useful. Have been using AirPods Pro as the mic and there's a low level fuzz that's in the recording, would want an easy way to clear that out.

Seconding this. It has an important side effect too: the tools used to machine-generate transcripts seem to dynamically adjust volume, so I am getting inaudible bleed-through (in my case from radio station interference!) being transcribed! I was able to solve this by preprocessing my file with a noise gate, but it might be wise for Descript to make some judgments about ignoring very, very quiet sounds.