Struct NativeAudioDecoder 

pub struct NativeAudioDecoder { /* private fields */ }

Streaming FFmpeg audio decoder normalized to project-rate planar f32.

Blocks keep the stream’s channels in its own order and name the speaker each feeds (NativeAudioDecoder::speakers). swresample turns each decoded frame into f32 at its own rate, rematrixing only a frame whose layout differs from the stream’s; the source timeline places it by its timestamp and alpha2_audio::Resampler, the converter a browser host runs as wasm, brings it to the project rate.

Implementations

impl NativeAudioDecoder

pub fn set_decode_policy(&mut self, policy: DecodePolicy)

Selects strict failure (default) or bounded, reported recovery.

pub fn set_decode_reporter( &mut self, reporter: impl Fn(DecodeDiagnostic) + Send + Sync + 'static, )

Receives each diagnostic before recovery continues, including before errors.

pub fn decode_diagnostics(&self) -> &DecodeDiagnostics

Cumulative corruption and concealment diagnostics, including seeks.

pub fn open( path: impl AsRef<Path>, project_sample_rate: u32, ) -> Result<Self, DecodeError>

Decodes the selected audio stream through the installed FFmpeg 8 runtime.

Stream selection and decoding use the linked FFmpeg libraries through rsmpeg; no command-line program is spawned. Sample-rate conversion is alpha2_audio::Resampler.

Errors

Returns a platform-neutral error for invalid rates, missing streams, unsupported layouts, codec failures, or malformed decoded audio.

pub fn set_block_frames(&mut self, frames: usize) -> Result<(), DecodeError>

Sets the maximum observable block size returned by decode_next.

Errors

Returns DecodeError::OutOfRange for zero frames.

pub const fn channels(&self) -> usize

Channels of every block this decoder returns: the source’s, 1 to MAX_AUDIO_CHANNELS.

pub const fn speakers(&self) -> SpeakerLayout

The speaker each plane of every block feeds: the layout the stream declares, in its own order, or the default for its channel count.

pub const fn speakers_assumed(&self) -> bool

Whether the stream declares no layout, so Self::speakers is the default for its channel count rather than what it says.

pub fn audio_duration(&self) -> Option<MediaTime>

Effective end of the selected audio source.

Containers commonly declare a slightly longer presentation duration than the selected stream because of codec priming and time-base rounding. Browser-authored projects use that container duration, so the adapter preserves it as the source boundary instead of rejecting the final presentation window.

Trait Implementations

impl Decoder for NativeAudioDecoder

fn streams(&self) -> &[StreamInfo]

Returns deterministic stream metadata and stream selection.

fn flush(&mut self) -> Result<(), DecodeError>

Flushes queued packets and frames without changing stream selection.

fn seek( &mut self, request: DecodeRequest, cancel: &CancelToken, ) -> Result<DecodeResult, DecodeError>

Seeks and returns what is presented at the request’s time.

fn decode_next( &mut self, cancel: &CancelToken, ) -> Result<DecodeEvent, DecodeError>

Advances packet draining by one observable event.