Scalable coding is the codec design pattern where one bitstream can be partially decoded to produce different quality, resolution, or frame-rate versions of the same content. A scalable encoder produces output where a "base layer" is fully decodable on its own — giving you, say, 480p at low bitrate — and one or more "enhancement layers" can be added to upgrade to 720p, 1080p or 4K at progressively higher bitrate. The same single file (or stream) serves viewers across a wide range of network and device capabilities, without re-encoding.
The idea has a long history. H.264 SVC (Scalable Video Coding, 2007) was the first widely-deployed version, used heavily in video conferencing and surveillance. HEVC SHVC (Scalable HEVC) extended the concept. AV1 has built-in scalability features for spatial and temporal layers. LCEVC (MPEG-5 Part 2) is the modern commercial implementation, designed as a codec-agnostic enhancement layer that works on top of any base codec.
For a product team in 2026, scalable coding has a niche but real role, distinct from adaptive-bitrate streaming (ABR). ABR stores multiple completely separate renditions of the same video at different qualities; scalable coding stores one file with embedded layers. ABR wins on simplicity and codec compatibility — every codec works, every player handles it, no special hardware needed. Scalable coding wins on storage efficiency (one layered file is smaller than three independent renditions) and on use cases where you genuinely need partial decoding (multi-device fan-out from a single contribution stream, real-time bandwidth adaptation in WebRTC, broadcast tier-down). Most commercial streaming uses ABR; scalable coding survives in WebRTC, broadcast, surveillance, and the growing LCEVC ecosystem.
