Hi there!
What you're trying to accomplish here is source separation in a music track. I will be accomplishing this task using an Artificial Neural Network Model, which will be trained upon preexisting data sets, and will be able to separate song stems. Like for example, the vocal stem, and the accompaniment stem (the backing track.) And hence the task put forth would be accomplished.
The processing will be happening on the user's side, and the newly separated audio would stream as soon as it has been processed. This processing can be optimized in development.
I would love to know any additional details, specifications about the project. Looking forward to delivering excellent high quality results!
Thanks.