Ability to filter just the lyrics

[wisack]

I noticed you are only able to filter the instrumentation or both, singing and instrumentation. It seems quite counter-intuitive as people could use just the singing stems for their own productions if you would enable the filtering of those. The underlaying song wouldn't matter that much either if you provided an ability to set the bpm of the song and generate multiple takes, it would create cool sampling capabilities. Nevertheless, even the filtering option would be great addition as you can timestretch and chop the sample.

I apologize if this doesn't quite concern the model development but this was the only discussion board I could find to leave an idea for the Suno platform.