Struct EncodedVideoTrack 

pub struct EncodedVideoTrack { /* private fields */ }

A video track that accepts pre-encoded frames directly, bypassing the libwebrtc software encoder pipeline entirely.

Obtain one via [PeerConnectionFactory::create_video_track_with_options] with TrackVideoEncoder::PreEncoded, then:

1. Attach EncodedVideoTrack::track() to a send-only transceiver.
2. Call push_encoded_frame whenever your encoder (VideoToolbox, NVENC, GStreamer, libvpx, …) produces a frame — at your rate, on any thread.

Timing

The WebRTC encoder thread is triggered internally each time you call push_encoded_frame, so you do not need to call push_video_frame separately. The dummy raw frame used to trigger it is cheap (pre-allocated; the I420 data is discarded by the encoder callback before it ever touches your encoded bytes).

Implementations

impl EncodedVideoTrack

pub fn on_encoder_feedback( &self, cb: impl FnMut(EncoderFeedback) + Send + 'static, )

Listen for encoder feedback for this track — EncoderFeedback::KeyFrameRequest (answer with an IDR promptly) and EncoderFeedback::RateUpdate (adapt your encoder’s target). Latest registration wins. The callback fires on WebRTC threads, like every other sink in this API.

pub fn track(&self) -> &VideoTrack

The underlying video track handle — pass track() (or the VideoTrack itself) to Transceiver::set_track; it derefs to &Track.

pub fn push_frame(&self, frame: EncodedVideoFrame) -> Result<()>

Inject a pre-encoded frame into the WebRTC RTP stack.

The call returns immediately; the frame is queued and forwarded to the RTP packetizer on the WebRTC encoder thread. Thread-safe — call from any thread, including a hardware encoder callback.

Set frame.width / frame.height to 0 to inherit from the track’s configured resolution — the PreEncodedOptions the track was created with.

pub fn push_frame_with_metadata( &self, frame: EncodedVideoFrame, user_data: &[u8], ) -> Result<()>

Inject a pre-encoded frame with per-frame metadata embedded as a protobuf trailer.

frame_id and capture_time_us are filled in internally, the latter from time_micros; the caller supplies only user_data. A capture time cannot be declared here the way it can on a raw Track: these frames arrive already encoded and the encoder clamps future capture timestamps, which is why this path correlates its metadata by FIFO order instead. Requires sender_metadata_transform to be called and the returned FrameTransform to be attached to the sender transceiver before the first SDP exchange.

Trait Implementations

impl Send for EncodedVideoTrack

impl Sync for EncodedVideoTrack