AV1 LTR Frame Acknowledgment Signaling

In real-time AV1 video streaming, receivers use transport-level feedback to confirm the receipt of Long-Term Reference (LTR) frames to maintain stream resilience without relying on bandwidth-heavy keyframes. This article explains how an AV1 receiver identifies an LTR frame, generates the appropriate RTP Control Protocol (RTCP) feedback messages, and communicates frame integrity back to the encoder to establish a shared reference state.

Identification via RTP Header Extensions

Before an acknowledgment can occur, the receiver must recognize that a decoded frame is intended as an LTR. In AV1 video transmission over RTP, this is handled through RTP header extensions:

RTCP Feedback Signaling

Once the receiver successfully decodes the designated frame and commits it to the designated reference buffer slot, it communicates this success back to the sender using RTCP:

Negative Acknowledgments and Loss Recovery

If an LTR frame is corrupted or dropped in transit:

Encoder-Side Synchronization

When the sender receives the RTCP acknowledgment containing the valid picture ID:

  1. Buffer Verification: The encoder marks the corresponding reference buffer as validated and synchronized with the remote receiver.
  2. Predictive Coding: In subsequent frames, the encoder can reference this confirmed LTR frame, ensuring that packet loss occurring on intermediate short-term reference frames does not cause decoding artifacts or freeze playback.