Skip to content

CONCEPT Cited by 1 source

Temporal-layer error resilience

Temporal layers (TL) are a form of temporal scalability in modern video codecs (including AV1) where the encoder organizes frames into a time-based hierarchy. The base layer (TL0) provides continuity at a lower frame rate, while enhancement layers (TL1, TL2, …) add intermediate frames for higher frame rates when conditions allow.

Error-resilience property

The key property: the base layer maintains decode continuity without relying on enhancement-layer frames. If enhancement-layer packets are lost or arrive too late, decoding proceeds using the base layer without stalling.

This enables differentiated protection: - FEC applied preferentially to base-layer data (protects continuity). - Enhancement-layer retransmissions handled conservatively — when RTT is low, retransmit; when RTT is high, skip without breaking decode flow.

Trade-off: resilience vs compression efficiency

Compared to a tightly-dependent prediction chain (each frame references the immediately preceding frame), a temporal-layer structure is less compression-efficient — the looser reference structure means weaker temporal correlation for prediction. Leaving TL enabled permanently can degrade quality at a given bitrate.

Adaptive activation

Meta's resolution (Source: sources/2026-06-22-meta-adopting-av1-for-real-time-communication-rtc-at-scale): enable TL adaptively. The sender monitors network feedback: - Loss rises → turn TL on (gain resilience). - Network recovers → turn TL off (regain compression efficiency).

This gives resilience when needed without sacrificing efficiency on stable networks. See patterns/adaptive-temporal-layer-activation.

Meta's structure

Meta uses a two-layer structure for AV1 RTC — base layer (TL0) at lower frame rate + one enhancement layer (TL1) at the target frame rate.

Seen in

Last updated · 559 distilled / 1,651 read