I use SDL_mixer, and I would echo your conclusion: it doesn't give enough low level access. It's fine for playing one background music track and doing fire & forget sound effects, but that's it. For my project, that's just barely kinda mostly sorta good enough for now, but the next time I revisit audio I'll be switching to either OpenAL or the raw ALSA API. I want to be able to change the volume of a triggered sound effect while it's playing. I want an immediate notification of when a sound has completed so I can sequence something else (non-audio) after it. I want to get feedback about how far through a sound I am. I want to do pitch bending. I wouldn't mind some 3D sound processing (although I think SDL_mixer can do some of that.)