shiguredo-audio-toolbox

Create an Encoder with EncoderConfig specifying codec AacLc, sample rate, channels, and optional bitrate, then call encode with interleaved i16 PCM and drain results via next_frame. Call finish() at the end to flush buffered samples.

Create a Decoder with DecoderConfig specifying the codec, input sample rate, and channels, then pass one compressed packet per decode call and read PCM via next_frame until it returns None. Output is always stereo 16-bit PCM at the input sample rate.

No, the crate is macOS-only because it binds directly to Apple's Audio Toolbox framework, and it emits a compile error on other platforms. Only documentation generation is possible elsewhere using DOCS_RS=1 cargo doc.

The decoder accepts exactly one packet per decode call, so you must drain the previous packet with next_frame until it returns Ok(None) before decoding again. Concatenating multiple packets into one call causes incorrect frame counts.

No, encoding supports only AAC-LC through EncoderCodec. FLAC is unavailable because AudioConverter does not support it, and HE-AAC or Opus encoding is not exposed even though decoding covers AAC-LC, MP3, and Opus.

No, both types are intentionally !Send because Apple does not document whether AudioConverter can move across threads. Use them on a single thread or serialize access yourself if cross-thread transfer is required.