Volume compressor / normalizer
Add an optional feature to enable (and adjust the degree of) volume compression so that the dynamic volume range is minimized within a given track.
And after that maybe add slow dynamic microphone based background noise level detection, and we have a new killer feature. Maybe split frequency band based or something, not all frqeuencies the same.
When in car or listening via motorcycle helmet, volume normalization would dramatically enhance the listening experience.
Yes, that would be useful. Audacity, an open source program, had such an effect (called normalisation); maybe you can use their code.
This would be very useful on podcasts with multiple voices where one is much quieter than another. If you listen while walking on city sidewalks during rush hour, it becomes impossible to hear all voices without blowing out your eardrums.