What Is Video Bitrate, and Why Does It Decide How Good a Video Looks?

Bitrate is how much data a video uses per second, usually written in kilobits or megabits per second. That is all it is. What makes it the most important number in video is what it represents: the budget an encoder has to describe everything happening on screen. Give it more and it can be faithful. Give it less and it has to start leaving things out.

Where the bits go

A video codec does not store each frame as a picture. It stores a complete frame occasionally, then describes what changed. A static shot of a wall changes almost nothing between frames, so it costs almost nothing to store. A handheld shot of rain on a busy street changes everything in every frame, and costs enormously more.

This is why the same bitrate produces wildly different results on different footage. Five megabits is luxurious for a slide presentation and thin for a football match at the same resolution. The encoder is not rationing arbitrarily; it is being asked to describe far more change per second.

Why a bitrate on its own means nothing

Asking whether 5 Mbps is enough is like asking whether a hundred pounds is a lot of money. For what? At 1280 x 720 and 30 frames a second, 5 Mbps is comfortable. At 3840 x 2160 the same bits are spread over nine times as many pixels and the picture will visibly fall apart on any complicated motion.

Dividing the bitrate by the pixels it has to cover gives a figure that survives comparison: bits per pixel per frame. As a rough guide for modern codecs, below about 0.05 you start seeing blocking on detailed motion, and above about 0.15 you are unlikely to see much return for the extra size. The compression estimator on this site works it out from a target size.

Starting figures, and where they come from

The honest answer to what bitrate should I use is that it depends, but that is useless as a starting point. So here is a table built from the rule above rather than from habit: each figure is the bitrate that lands around 0.10 bits per pixel per frame for H.264, which is comfortable for ordinary footage, and around 0.06 for the newer codecs that need fewer bits to reach the same result.

Double every figure for 60 frames a second, because you are describing twice as many pictures. Then adjust for the footage: a talking head or a screen recording sits happily at half these numbers, and fast sport, heavy grain or confetti wants half again on top. Treat the table as the middle of a range, not a specification.

Uploading somewhere changes the calculation entirely. YouTube, Instagram and the rest re-encode everything they receive, so the bitrate a viewer eventually sees is theirs and not yours. Give them more than they need, within whatever upload limit applies, because their encoder can only work with what arrives.

How to read the bitrate of a file you already have

Two numbers are usually reported and they are not the same thing. Overall bitrate covers everything in the file including audio; video bitrate covers the picture alone. Audio is commonly 128 to 320 kbps, so on a small file the difference is worth noticing and on a large one it is noise.

The video file inspector on this site reads both, along with the resolution, frame rate and codec, which is exactly the set of numbers this article says you need before a bitrate means anything. It reads the file in the browser, so nothing is uploaded anywhere.

Constant, variable and constrained

• Constant bitrate holds the same rate throughout, wasting bits on simple scenes and starving complex ones. It exists because some broadcast and streaming systems need a predictable rate, not because it looks good.
• Variable bitrate lets the encoder spend where it matters. Better quality at the same average, and the standard choice for files.
• Constrained variable bitrate allows variation within a ceiling, which is what live streaming needs: it can adapt, but it cannot exceed what the connection will carry.

Targeting quality instead

Most encoders offer a quality target as an alternative, commonly a CRF value. Instead of naming a bitrate, you name a standard, and the encoder spends whatever each scene needs to meet it. The result is usually better looking and often smaller than a fixed bitrate at the same average.

The trade is predictability: you find out the file size afterwards. When a hard size limit applies, target the bitrate. When you want the best file for the least size, target quality.

Newer codecs change what a number means

H.265 and AV1 generally reach the same visible quality at roughly half to two thirds the bitrate of H.264. A figure that looks thin for one codec can be perfectly adequate for another, which is another reason a bitrate quoted without its codec and resolution is not information.

Practical rules

1. Never judge a bitrate without knowing the resolution, the frame rate and the codec.
2. If the bitrate is tight, lower the resolution rather than accepting a starved picture. Fewer pixels described well beats more described badly.
3. For files you keep, prefer a quality target and a slower encoder preset.
4. For a size limit, work backwards from the limit and leave a few per cent of headroom.