Model to remove traffic noise? #77
Description
Is your feature request related to a problem? Please describe.
A classic example of a problem recording audio is traffic(cars, trains, planes etc.). I could be interesting to see to what extent a DL model could take care of that. Ex material with traffic in the recording: https://www.youtube.com/watch?v=hb8yQO_MDdI (some minutes in)
Describe the solution you'd like
A model trained on removing traffic noise.
Describe alternatives you've considered
I've tried the included voice separating modules, but I guess they have been trained on music?
Additional context
Why deal with audio for video? Because the Blender video editor can roundtrip audio in Audacity: https://github.com/tin2tin/audacity_tools_for_blender