SVC (Scalable Video Coding) is a video encoding technique that represents a single video signal as a base layer plus one or more enhancement layers, all packed into a single bitstream; the SFU can drop enhancement layers in real time to deliver a lower-bitrate version to a receiver without asking the sender to re-encode. The base layer alone is decodable and produces a low-quality picture; each additional spatial, temporal, or quality layer refines it. In the WebRTC context, VP9 with SVC and AV1 with its native SVC support are the most relevant codecs; H.264 does not support standard SVC in browsers. The main advantage over simulcast is efficiency: the sender transmits a single encoded stream rather than multiple independent streams, which reduces the sender's upload bandwidth and encoding CPU. The SFU can perform layer switching — dropping or adding layers — in response to receiver bandwidth changes without requiring a keyframe from the sender, which means quality adaptations are smoother and faster than with simulcast. The practical gotcha is codec compatibility: VP9 SVC requires the SFU to understand the VP9 payload structure to split layers correctly, and not all SFU implementations support it well. For large live classrooms where sender bandwidth is a constraint — for example participants on mobile data — SVC can meaningfully reduce upload requirements compared to simulcast while retaining fine-grained quality adaptation at the server.
