Module audio::s24

S24: one packed 24-bit audio sample.

AUDIO_FORMAT_I24 buffers hold signed 24-bit samples packed in 3 bytes, little-endian (no padding byte). S24 is exactly that layout — sizeof(struct S24) == 3, alignment 1 — so a packed buffer is a []S24:

[]S24 out = io.output_i24();             // frames * channels samples
out[i] = S24::from_f32(0.5f);
int raw = out[i].value();                // -8388608 .. 8388607

and audio_convert.h's audio_f32_to_i24 / audio_i24_to_f32 convert whole slices of them. Freestanding-safe.