Is there knowledge regarding how efficient the library is compared to other audio libraries?

[mavavilj]

Is there knowledge regarding how efficient the library is compared to other audio libraries?

[mackron]

Moved to the discussions section.

I'm not aware of any community benchmarks and I've not done a lot of optimization work. In general, the code is written with good practice in mind such as avoiding redundant data copying and whatnot. There's no grievous inefficiencies or anything. The decoders are well optimized (compile with SSE2 enabled for more performance).

[mavavilj]

I was thinking of doing such benchmark, but I thought that if it's known already that some library is written using best knowledge, then there wouldn't be much to compare really. But OTOH I don't expect this and e.g. PortAudio to have similar latency (I expect MiniAudio to be faster, since it's smaller). Thus it might be useful for someone to know, why they shouldn't pick PortAudio or something for their use case.

[mavavilj]

But OTOH it'd be a shame, if a library did in fact contain an inefficiency (compared to another library), which could be eliminated for better throughput latency. That's why I was interested in asking this question.

[mackron]

In terms of latency, the general feedback I've heard has been positive. I don't actually recall anybody complaining about latency (had an issue where I was misusing PulseAudio once, but that's fixed). But again, I'm not aware of any actual benchmarking so can't say for sure how it compares to others. Understand your concern about the possibility of some hidden inefficiency that hasn't yet been detected, and it's a fair point.

I do want to do an optimization pass at some point, but I'm more concerned about getting the library feature complete and stabilized. It's getting there slowly.

If you do indeed do some benchmarking, I'd be happy to give you some advice on what's going on in the library.