Audio functionality for the SDL library.
All audio in SDL3 revolves around SDL_AudioStream. Whether you want to play or record audio, convert it, stream it, buffer it, or mix it, you're going to be passing it through an audio stream.
Audio streams are quite flexible; they can accept any amount of data at a time, in any supported format, and output it as needed in any other format, even if the data format changes on either side halfway through.
An app opens an audio device and binds any number of audio streams to it, feeding more data to it as available. When the devices needs more data, it will pull it from all bound streams and mix them together for playback.
Audio streams can also use an app-provided callback to supply data on-demand, which maps pretty closely to the SDL2 audio model.
SDL also provides a simple .WAV loader in SDL_LoadWAV (and SDL_LoadWAV_IO if you aren't reading from a file) as a basic means to load sound data into your program.
Audio data passing through SDL is uncompressed PCM data, interleaved. One can provide their own decompression through an MP3, etc, decoder, but SDL does not provide this directly. Each interleaved channel of data is meant to be in a specific order.
Abbreviations:
These are listed in the order they are laid out in memory, so "FL, FR" means "the front left speaker is laid out in memory first, then the front right, then it repeats for the next audio frame".
This is the same order as DirectSound expects, but applied to all platforms; SDL will swizzle the channels as necessary if a platform expects something different.
SDL_AudioStream can also be provided channel maps to change this ordering to whatever is necessary, in other audio processing scenarios.