AV1 Operating Points and Scalable Bitstreams

In the AV1 video codec, operating points serve as standardized identifiers that define specific combinations of spatial and temporal layers within a single scalable bitstream. They provide the necessary metadata for decoders, media servers, and extractors to isolate and parse a fully compliant sub-bitstream tailored to specific display capabilities, bandwidth constraints, or processing limitations without requiring re-encoding. By indexing target resolutions, frame rates, and decoding complexity levels, operating points make selective sub-bitstream extraction efficient and deterministic.

Scalability in AV1

AV1 natively supports Scalable Video Coding (SVC), allowing a single video stream to contain multiple overlapping layers:

Because a full scalable AV1 bitstream contains interdependent data for multiple visual presentations, downstream systems need a standardized way to decouple these layers.

The Role of Operating Points

An operating point in AV1 is represented by the 12-bit syntax element operating_point_idc located within the Sequence Header Open Bitstream Unit (OBU). It specifies:

  1. Target Layers: A bitmask indicating which temporal and spatial layers must be decoded to produce a specific visual output.
  2. Decoder Conformance: The specific profile, tier, and level required to decode that specific subset of layers.
  3. Initial Display Delay: The buffer parameters and latency characteristics specific to that presentation layer.

The Extraction Process

When a client or a streaming server extracts a sub-bitstream, the operating point dictates the extraction workflow:

Practical Applications

Operating points streamline adaptive media delivery across multiple scenarios: