How to choose a bitrate

Bitrate is how much data a video uses per second of playback, usually measured in megabits per second (Mbps). It is the single biggest factor in both file size and visual quality: too low, and compression artifacts show up as blocking or smearing; too high, and you're spending storage and bandwidth on data the eye can't actually use.

The three common ways to control it

CBR — Constant Bitrate

The encoder targets the same bitrate for every second of the video, regardless of how complex the scene is. This makes file size and bandwidth use perfectly predictable, which is why it's the default for live broadcast and many streaming pipelines — but it wastes data on simple scenes and can struggle on complex ones.

VBR — Variable Bitrate

The encoder spends more bits on complex, high-motion scenes and fewer on static or simple ones, aiming for a consistent quality level rather than a consistent data rate. Average file size is usually smaller than CBR at the same perceived quality, at the cost of a less predictable bitrate over time.

CRF — Constant Rate Factor

Common in offline encoding (not typically used for live streaming). Instead of specifying a bitrate at all, you specify a quality level, and the encoder uses however many bits it needs, scene by scene, to hit it. This usually gives the best quality-to-size ratio when file size isn't tightly constrained, but the resulting file size can't be predicted exactly in advance.

What actually drives how many bits you need

• Resolution. More pixels need more data to describe at the same quality.
• Frame rate. More frames per second means more information to encode per second of playback.
• Motion and detail. A static talking-head shot compresses far more easily than fireworks, fast camera pans, or fine textures like grass or water.
• Codec. As covered on the H.264 vs H.265 page, a more efficient codec needs fewer bits to reach the same quality.

Practical starting ranges

These are reasonable starting points for general-purpose, moderate-motion content (adjust upward for high-motion sports or gaming footage, and downward for mostly-static content):

Audio is a much smaller share of the total but still worth setting deliberately: 128 kbps is a reasonable default for stereo AAC or Opus, 192–256 kbps for music-heavy content where clarity matters more, and as low as 64–96 kbps for spoken-word content like commentary or narration.

A simple way to estimate file size

File size in megabytes is roughly:

(video bitrate + audio bitrate, in Mbps) × duration in seconds ÷ 8

For example, a 10-minute (600-second) video encoded at 5 Mbps video plus 128 kbps (0.128 Mbps) audio works out to roughly (5 + 0.128) × 600 ÷ 8 ≈ 384 MB.

When in doubt

If the output is a one-off file rather than a live stream, CRF-based encoding generally gives the best result for the least effort — pick a quality level, not a bitrate, and let the encoder allocate data where the content actually needs it. Reach for CBR only when a predictable, constant data rate genuinely matters, such as live broadcast or a strict bandwidth cap. See adaptive streaming for how bitrate choices extend into serving multiple quality levels of the same video at once.