US What bitrate means
Bitrate, also written bit rate, is the number of bits carried or processed each second. It is written in bits per second: kilobits (kb/s), megabits (Mb/s) or gigabits (Gb/s). A lower-case b means bits and a capital B means bytes, and one byte is eight bits, which is the source of a lot of confusion between network speeds and file sizes.
The term applies to any digital link, but it matters most in audio and video. There, bitrate is the budget the encoder has for describing each second of sound or picture. Give it more bits and it keeps more detail; give it fewer and it has to discard more, which eventually shows as blocking, smearing and banding.
Bitrate also decides two practical numbers: how much bandwidth a stream needs to travel over a network, and how much storage a recording fills.
What decides video bitrate
No single bitrate is right for all video. The rate needed for a clean picture depends on several factors:
-
//
Resolution. A 4K frame has four times the pixels of 1080p, so it needs several times the bitrate, though not usually a full four times.
-
//
Frame rate. Doubling from 30 to 60 frames per second adds data, typically less than double because consecutive frames are similar.
-
//
Codec. H.265 typically needs around half the bitrate of H.264 for similar quality, and AV1 and H.266 go lower still.
-
//
Motion and detail. A static camera on a quiet corridor needs far fewer bits than a busy road, sports or a camera on a moving vehicle.
-
//
Bit depth and color sampling. 10-bit and 4:2:2 video carry more information than 8-bit 4:2:0, which raises both raw and compressed rates.
Typical video bitrates
The gap between raw and compressed video shows why codecs exist. Uncompressed video needs a dedicated cable such as SDI, while compressed streams fit easily on a network.
| Format | Uncompressed (10-bit 4:2:2) | Typical H.264 | Typical H.265 |
|---|---|---|---|
| 720p30 | About 553 Mb/s | 2–4 Mb/s | 1–2 Mb/s |
| 1080p30 | About 1.24 Gb/s | 4–8 Mb/s | 2–4 Mb/s |
| 1080p60 | About 2.49 Gb/s | 6–12 Mb/s | 3–6 Mb/s |
| 2160p30 (4K) | About 4.98 Gb/s | 15–30 Mb/s | 8–15 Mb/s |
| 2160p60 (4K) | About 9.95 Gb/s | 25–50 Mb/s | 13–25 Mb/s |
Uncompressed figures are for active picture only. The compressed ranges are typical for live streaming and recording and vary with content and encoder settings. The raw rates are why 3G-SDI carries 1080p60 and 12G-SDI is needed for 4K60, and why SDI capture cards bring uncompressed video into a computer for analysis without compression artefacts.
CBR vs VBR
Encoders use rate control to decide how many bits each frame gets. The two main modes behave quite differently:
-
//
Constant bitrate (CBR). The encoder holds a steady rate whatever the content. Bandwidth is predictable, which suits live streaming and fixed links, but simple scenes waste bits and complex ones may lose quality.
-
//
Variable bitrate (VBR). The rate rises for complex scenes and falls for simple ones, giving better quality for the same average size. It suits recording and on-demand video.
-
//
Capped VBR. VBR with a maximum, a common compromise for streaming over links that cannot take large spikes.
-
//
Constant quality. Modes such as CRF or fixed QP target a quality level and let the bitrate follow, useful when file size matters less than consistent quality.
Bitrate in surveillance and multi-camera systems
In systems with many cameras, bitrate becomes a sizing exercise. Total bandwidth is the sum of every stream, and storage is that total multiplied by the retention period. Thirty-two 1080p cameras at 4 Mb/s each produce 128 Mb/s, which is roughly 1.4 TB a day, so the choice of codec and rate control directly sets the disk capacity an NVR that records IP cameras needs.
For live distribution, hardware video encoders let you set the bitrate, rate control mode and stream count per input, so a feed can be sized to the network it has to cross.
Bitrate also affects video analytics. Heavy compression blurs fine detail such as number plates, small defects or distant people, and AI models trained on clean images can miss them. Where accuracy matters, run analysis on the camera’s full-quality stream or on uncompressed capture, and record a lower-bitrate copy for storage.
Quick answers on bitrate
Bitrate: common questions
Does a higher bitrate mean better quality?
Up to a point. With the same codec and resolution, more bits keep more detail and fewer artefacts. Beyond a certain level the gain becomes invisible and only the file size grows.
What bitrate should I stream at?
Match it to the resolution, codec and upload speed. As a rough guide, 1080p60 in H.264 needs around 6–12 Mb/s, and the stream should use no more than about 70–80% of the available upload bandwidth.
How do I work out file size from bitrate?
Multiply by time and divide by eight. Bitrate is in bits and file size in bytes. A stream at 8 Mb/s for one hour is 8 × 3,600 ÷ 8 = 3,600 MB, or about 3.6 GB.
What is the difference between bitrate and bandwidth?
One is used, the other is available. Bitrate is the data a stream needs per second. Bandwidth is the capacity of the link carrying it. The total bitrate of all streams must fit within the bandwidth, with headroom.
What is a good audio bitrate?
Around 128–320 kb/s for lossy music. AAC at 128 kb/s or more suits most uses. Speech can go much lower, and uncompressed CD audio runs at 1,411 kb/s.
We fit capture cards and configure encoding in industrial and edge AI computers as standard, sizing bitrate, codec and storage to the number of streams before the system ships.