There is a movie I like and it has unreleased soundtrack. Because of the way each thing is mixed, it is possible to separate the voices, music and sound effects like car engines using some program. People have done it before, however, what program should I use? Thank you.
You are talking about some sophisticated signal analysis. It would be similar to what the karaoke people do to remove vocals from some songs. I think it starts with Fourier-based filtering and probably involves some manual manipulation in the time domain. These days there are probably some AI tools that can help. For those, I would start by reviewing the plethora of available R packages. The Python community might be a better place to look, but I'm not a Python person, so I would start with R.