This could save many people’s lives.
A plugin that would let me point on a few sample “Umm… Emm… Hmmm…” sounds on a voice track, and would detect similar sound on the whole track.
Then, it should (1) mute/trim all or (2) mark them (3) or let hear them one after another and decide - mute / trim / keep.
Ideally, it should be collecting a dataset from the results, to grow an intelligence to identify them even without providing a sample, eventually.
In fact, that could work for any sound. I could ask to remove the f word from a long podcast for instance. Could work as a censorship tool as well, depending on the samples I provide.
Thanks and will be waiting for such tool to happen!