Last time I used Rust to build a music player, the resulting audio quality was much worse than my C++ version. I'm sure these new kids on the block ("Rustaceans" or whatever they call themselves) know what they are doing when building web applications and CLIs, but when audio quality has to 120% of the original sound, nothing beats true and tested C++. I cannot explain how or why, but programs made with C++ just sounds better, fuller and more authentic than programs made with other languages.
C++ has a lower-latency linker.
But Rust's advanced type system will surely result in better instrument separation and wider scene.
The borrow checker acts as a subtle low-pass filter, rolling off some of the high end and leading to less clarity.
This is because rust's borrow checker is incapable of representing some perfectly valid sounds.
Have you tried using an audio player written in hand-rolled assembly?
It sounds like a good idea but it produces worse results on other CPUs. It's best to always compile on every install as it uses the best instruction set extensions available. AVX2 produces a nice musky sound. But only with AVX-512 do you get rich aftersounds.
Of course the solution is not portable. It works for me, though. Ever since I spent $50k on my gold-plated, custom-tuned 486dx in 1992 I haven't been able to justify an upgrade (many of the Pentium series chips produced terrible audio, even in the high-end market).
Due to Rust's fearless concurrency, every time I have to add a new instrument, I either have to make a copy of the entire track to add it, or temporarily borrow the bass from the guitar track, which leads to tiny jumps in the audio. Notwithstanding the need for unsafe when I want to put the amp to 11.