Ryan C. Gordon
Ryan C. Gordon
So in the time since this originally came up, SDL3's SDL_AudioStream came along and is _super_ powerful for this; it can change both input and output sample rate on the...
I think it's too much complexity that most things aren't going to use. The exception is talking to both IPv4 and IPv6 at the same time, and the usual solution...
Hmm...@twolife, can you verify? We could maybe add a quirk to force vsync or just make SDL_GL_SwapBuffers stall to keep to ~60fps.
We should probably do both, vsync and _also_ wait, so it looks smoother and also doesn't iterate too quickly if the refresh rate is too high. But maybe that's silly,...
This is likely a leftover from when this was an HLSL preprocessor (which does some questionable things to the output like this). If the rest of the tests still pass...
Generally alignment stuff is because the SSE bits were written in an era where it was considered disastrous to attempt unaligned accesses. My understanding is modern Intel CPUs don't really...
I think the shift was trying to preserve the most significant integer bits, but I'm not really sure if this was wise or not. If you produce something with similar...
> does this change fix the audio_resampleLoss test? That bug is something I introduced in the latest round of resampler work, I'm working on it.
> While above is our program's flaw I really don't think it is, and where MojoAL has a problem with this, I consider it a MojoAL bug.
Oh, that's a good idea!