Simulcast

Simulcast is a WebRTC sender-side technique where a single camera or screen-share is encoded into two or three independent streams — typically a high, medium, and low resolution — and all of them are uploaded to the SFU simultaneously. The SFU then selects which layer to forward to each subscriber based on the subscriber's available download bandwidth and the number of visible participant tiles on their screen. A viewer on a slow connection receives only the low-resolution layer; a viewer on a strong connection receives the high-resolution layer; and the sender's upload cost is fixed regardless of how many viewers there are. This is the most widely deployed mechanism for adaptive quality in WebRTC live classes because it requires no transcoding on the server — the SFU is simply choosing among already-encoded packets. The trade-off is that the sender must encode and upload multiple streams at once, increasing their upload bandwidth requirement by roughly 1.5–2x compared to a single stream. Simulcast is distinct from SVC (Scalable Video Coding): with simulcast each layer is an independent H.264 or VP8 stream that can be decoded alone, whereas SVC layers are interdependent within a single encoded bitstream. Browser support for simulcast is solid across Chrome and Firefox via the Unified Plan SDP API; the platform's signaling layer controls which layers the sender activates and which the SFU forwards.

The handbook · 7 volumes

Fora Soft Video Engineering Handbook

Simulcast lets each student get a stream that fits their connection without the teacher re-encoding for everyone. Our seven-volume handbook covers the whole video stack; for simulcast start with Volume 3 on WebRTC and choosing the right layer per receiver, with encoder settings in Volume 1 and congestion control in Volume 2. Written by Nikolay Sapunov, CEO at Fora Soft.

Browse the handbook on Amazon →7 volumes · 9,000+ pages · Kindle & paperback