I'm unsure if I've already submitted a request regarding this matter.

I viewed a YouTube video showcasing Sony's "diff-a-riff" technology, seemingly an early, unreleased audio and sound competitor product, yet already quite impressive. The system allows users to input their audio and specify additions or modifications via text prompts. For example, if I have a piano melody and wish to add prominent 808 bass, I can input the request and receive the modified audio. However, while effective for simple additions, applying it to a full remix results in unintended, widespread changes to the entire song. Reducing the remix intensity yields no changes. Precise control over additions and subtractions remains challenging. I trust this explanation clarifies my concerns.