How to use SDL_MixAudioFormat in SDL3?

[huangjinshe]

Looks like the SDL_MixAudioFormat document still not update.
it still using Uint8 * type for stream, but SDL3 already using SDL_AudioStream* for stream in callback.

[huangjinshe]

We usally use this for change audio volume before in SDL2.

[icculus]

SDL_MixAudioFormat still operates on an array of audio data; when using the callbacks in SDL3, where you might have previously done something like this:

static void SDLCALL MyAudioCallback(void *userdata, Uint8 * stream, int len)
{
     // I have audio data in a global, MyAudioBuffer, that I want to adjust and copy into stream...
     // you have to initialize `stream` to silence first, because we're mixing into it!
     SDL_memset(stream, 0, len);
     SDL_MixAudioFormat(stream, MyAudioBuffer, MyAudioBufferFormat, len, MyVolume);
}

...now you have to put that data into an SDL_AudioStream instead of copying to an array:

static void SDLCALL MyAudioCallback(void *userdata, SDL_AudioStream * stream, int len)
{
     // I have audio data in a global, MyAudioBuffer, that I want to adjust and copy into stream...
    Uint8 *buffer = (Uint8 *) SDL_calloc(1, len);  // SDL_calloc zeroes out the buffer after allocating.
    if (buffer) {
        SDL_MixAudioFormat(buffer, MyAudioBuffer, MyAudioBufferFormat, len, MyVolume);
        // okay, _now_ we're ready to put the mixed data into the stream!
        SDL_PutAudioStreamData(stream, buffer, len);
        SDL_free(buffer);
    }
}

This doesn't necessarily have to use SDL_calloc, but it needs a buffer somewhere it can mix to before putting it in the stream.

[ell1e]

Hm sads, I wish there was a way to avoid more buffer copies.