Skip to content

CONCEPT Cited by 1 source

Constant bitrate (CBR)

Definition

Constant bitrate (CBR) is a video-encoder rate-control mode in which the encoder targets a fixed bytes/sec output, regardless of scene complexity. Perceived quality floats scene-by-scene: easy content gets over-spent, hard content gets under-spent.

Opposite of VBR, which targets constant quality and lets bytes/sec float.

Why CBR was the default

CBR dominated early streaming-pipeline design because its predictability made the rest of the stack simple (Source: sources/2026-04-02-netflix-smarter-live-streaming-vbr-at-scale):

  • Capacity planning is trivial. "If a server could safely deliver around 100 Gbps of Live traffic, and each stream averaged close to its nominal rate, we could admit on the order of twenty thousand concurrent sessions per server and be confident we were operating within limits."
  • Admission control is trivial. Current Gbps on the wire is a faithful proxy for how full the server is: "traffic stayed relatively flat. The network saw a smooth, easy-to-reason-about load profile, and large changes in throughput almost always reflected a real change in usage, not just a different scene on screen."
  • Player ABR logic is well-matched. Nominal bitrate labels on ladder rungs correspond to actual second-by-second bytes/sec, so the throughput-vs-label comparison the ABR algorithm performs is meaningful.

Why CBR loses efficiency

Real content is non-uniformly complex. Under CBR, "in simple scenes, we spend more bits than we need; in complex scenes, we sometimes don't spend enough." Netflix's WWE RAW worked example (Source: sources/2026-04-02-netflix-smarter-live-streaming-vbr-at-scale) makes the inefficiency visible: with CBR at 8 Mbps nominal, the bitrate stays clustered around 8 Mbps whether the scene is the static "waiting room" or the confetti-filled shot. The two scenes have wildly different compression costs at matched quality, but CBR spends the same bit budget on both.

Netflix's CBR-era Live admission model

Pre-2026-01-26, Netflix Live ran CBR with the following implicit assumptions (Source: sources/2026-04-02-netflix-smarter-live-streaming-vbr-at-scale):

  • "For each encoded stream, we configured a resolution and a nominal bitrate — for example, a 1080p stream targeting 5 Mbps — and the actual bitrate stayed close to that target over time."
  • Per-server session admission based on current aggregate traffic + headroom.
  • The total traffic on a server "would change mainly when members joined or left; as long as concurrency was stable, traffic stayed relatively flat."

When Netflix cut over to capped VBR on 2026-01-26, both the per-stream predictability and the admission-control invariant changed, driving the need for the reserve- against-nominal fix.

Quality side of the CBR↔VBR comparison

Quality-at-matched-bitrate is not the right comparison axis. The right comparison is matched-quality bitrate: a VBR stream at the same perceptual quality as a CBR stream uses fewer bits on average. Netflix measured this via VMAF rung-by-rung: "wherever VBR fell more than about one VMAF point below CBR, we increased its nominal bitrate just enough to close the gap" (Source: sources/2026-04-02-netflix-smarter-live-streaming-vbr-at-scale).

Seen in

Last updated · 319 distilled / 1,201 read