A question about constructing high-resolution music waves

[gogo9th]

Hi,

I am working on implementing an automated system that enhances the sound quality of real-time music streams generated by lossy compression (e.g., MP3, AAC). I find your open source project to be very valuable to my current research direction. 

I wonder if your technique can be universally applied to converting lossy-compression-generated musical waves to a lossless form in real time. The challenges are:

1. There are many genres of music, so I am not sure if it's okay to apply the same model to any kind of sound waves

2. Real-time sound processing is sensitive to delay, so the processing time of running inputs through a model should be short

Also, I wonder if you are aware of any other related open-source project regarding this research problem of constructing high resolution music waves from lossy compression in a universal manner (i.e., applicable to any genres & any compression methods of music waves). It was hard for me to find any other good projects besides your work from online.. 

I appreciate your feedback.