CONCEPT Cited by 1 source
Asymmetric codec design¶
Asymmetric codec design is the RTC architecture where the two sides of a video call use different codecs for sending vs receiving, exploiting the fundamental asymmetry between encoding complexity (high) and decoding complexity (lower).
Mechanism¶
In Meta's AV1 RTC deployment: mid-range devices that cannot encode AV1 in real-time often can decode AV1 in real-time. This enables:
- High-end device → mid-range device: high-end encodes and sends AV1; mid-range decodes AV1 (receives better quality).
- Mid-range device → high-end device: mid-range encodes and sends H.264/AVC (lower complexity); high-end decodes H.264.
The result: the mid-range device benefits from AV1's compression efficiency on its incoming video stream without bearing the encoding cost (Source: sources/2026-06-22-meta-adopting-av1-for-real-time-communication-rtc-at-scale).
Why it matters¶
Asymmetric codec design significantly increases AV1 coverage across the device fleet. Without it, AV1 benefits are limited to calls where both parties can encode AV1. With it, any call where at least one party has a high-end device delivers AV1-quality video in at least one direction.
Prerequisites¶
- Both codecs negotiated at call setup.
- Continuous feedback of encoding/decoding latency between peers.
- The concepts/codec-complexity-adaptation mechanism to detect and enforce the asymmetry dynamically.
Seen in¶
- sources/2026-06-22-meta-adopting-av1-for-real-time-communication-rtc-at-scale — canonical instance of asymmetric codec design for fleet-wide AV1 coverage expansion.
Related¶
- concepts/device-eligibility-ml — determines which devices are decode-only eligible
- concepts/codec-complexity-adaptation — the runtime mechanism that detects and enforces asymmetry
- patterns/asymmetric-send-receive-codec — the pattern page