Audio is hard in general and emulation probably even more so. The main issue is that for games you want as low latency as possible, when you control things imm

Audio is hard in general and emulation probably even more so. The main issue is that for games you want as low latency as possible, when you control things immediately it's easy to know when actual triggers happens, but for emulation you might need to copy or snapshot the entire state whilst under a strict timing deadline. The options are, bigger buffers (easier to emulate, more time for things.. worse latency) or smaller ones (less latency, but timing is harder and if there is extra work that has to fit within the deadline).