RTCEncodedVideoFrame

Limited availability

This feature is not Baseline because it does not work in some of the most widely-used browsers.

The RTCEncodedVideoFrame of the WebRTC API represents an encoded video frame in the WebRTC receiver or sender pipeline, which may be modified using a WebRTC Encoded Transform.

Note: This feature is available in Dedicated Web Workers.

Instance properties

RTCEncodedVideoFrame.type Read only

Returns whether the current frame is a key frame, delta frame, or empty frame.

RTCEncodedVideoFrame.timestamp Read only

Returns the timestamp at which sampling of the frame started.

RTCEncodedVideoFrame.data

Return a buffer containing the encoded frame data.

Instance methods

RTCEncodedVideoFrame.getMetadata()

Returns the metadata associated with the frame.

Description

Raw video data is generated as a sequence of frames, where each frame is a 2 dimensional array of pixel values. Video encoders transform this raw input into a compressed representation of the original for transmission and storage. A common approach is to send "key frames" that contain enough information to reproduce a whole image at a relatively low rate, and between key frames to send many much smaller "delta frames" that just encode the changes since the previous frame.

There are many different codecs, such as H.264, VP8, and VP9, each that have a different encoding processes and configuration, which offer different trade-offs between compression efficiency and video quality.

The RTCEncodedVideoFrame represents a single frame encoded with a particular video encoder. The type property indicates whether the frame is a "key" or "delta" frame, and you can use the getMetadata() method to get other details about the encoding method. The data property provides access to the encoded image data for the frame, which can then be modified ("transformed") when frames are sent or received.

Examples

This code snippet shows a handler for the rtctransform event in a Worker that implements a TransformStream, and pipes encoded frames through it from the event.transformer.readable to event.transformer.writable (event.transformer is a RTCRtpScriptTransformer, the worker-side counterpart of RTCRtpScriptTransform).

If the tranformer is inserted into a video stream, the transform() method is called with a RTCEncodedVideoFrame whenever a new frame is enqueued on event.transformer.readable. The transform() method shows how this might be read, modified by inverting the bits, and then enqueued on the controller (this ultimately pipes it through to the event.transformer.writable, and then back into the WebRTC pipline).

js
addEventListener("rtctransform", (event) => {
  const async transform = new TransformStream({
    async transform(encodedFrame, controller) {
      // Reconstruct the original frame.
      const view = new DataView(encodedFrame.data);

      // Construct a new buffer
      const newData = new ArrayBuffer(encodedFrame.data.byteLength);
      const newView = new DataView(newData);

      // Negate all bits in the incoming frame
      for (let i = 0; i < encodedFrame.data.byteLength; ++i) {
        newView.setInt8(i, ~view.getInt8(i));
      }

      encodedFrame.data = newData;
      controller.enqueue(encodedFrame);
    },
  });
  event.transformer.readable
    .pipeThrough(transform)
    .pipeTo(event.transformer.writable);
});

Note that more complete examples are provided in Using WebRTC Encoded Transforms.

Specifications

Specification
WebRTC Encoded Transform
# rtcencodedvideoframe

Browser compatibility

BCD tables only load in the browser

See also