Video Frame Size Calculator (uncompressed MB per frame & data rate)

Understanding the video frame size calculator — concept and principle

How big is one frame of uncompressed video, and how much data does it produce per second? This calculator works out the size of a frame from resolution, chroma subsampling (4:4:4, 4:2:2 or 4:2:0) and bit depth, then the data rate and storage needed for a recording. It is useful for planning capture cards, recorders, interfaces and storage.

4:2:2 keeps half the colour samples, averaging 20 bits per pixel at 10-bit depth.

How it is calculated

Luma samples = width × height

Chroma samples per component: 4:4:4 → width × height; 4:2:2 → (width ÷ 2) × height; 4:2:0 → (width ÷ 2) × (height ÷ 2)

Frame size (bits) = (luma + 2 × chroma) × bit depth; bytes = bits ÷ 8

Bits per pixel = frame bits ÷ (width × height)

Data rate = frame bits × frame rate; storage = data rate × time ÷ 8

MB = 10⁶ bytes; MiB = 2²⁰ bytes.

The principle

Digital video stores brightness (luma, Y′) for every pixel and two colour-difference signals (Cb, Cr). Human vision is less sensitive to fine colour detail than to brightness detail, so colour can be stored at lower resolution. 4:2:2 halves the colour samples horizontally and 4:2:0 halves them both ways, giving on average 2 or 1.5 samples per pixel instead of 3. Multiplying samples by bits per sample gives frame size, and multiplying by frames per second gives the data rate.

Worked example

Default inputs: 3,840 × 2,160, 4:2:2, 10-bit, 25 fps, 10 minutes.

Luma = 3,840 × 2,160 = 8,294,400 samples.
Chroma = 2 × (1,920 × 2,160) = 8,294,400 samples, so 16,588,800 samples in total.
Frame = 16,588,800 × 10 = 165,888,000 bits = 20.74 MB = 19.78 MiB; 20 bits per pixel.
Data rate = 165,888,000 × 25 = 4.147 Gbit/s.
Storage = 31.1 GB per minute, 311 GB for 10 minutes.

Each point is the calculator re-run with only “Width (px)” changed; the gold point is the default example. Change the inputs above and this chart follows your numbers.

Width (px)	Size of one frame (MB)
960	5.18
1,920	10.37
2,880	15.55
3,840	20.74
4,800	25.92
5,760	31.1
7,680	41.47

How to read the result

These numbers show why uncompressed 4K video needs fast interfaces and large storage, and why cameras normally record compressed formats such as H.264, HEVC or ProRes, which are far smaller. The figure counts only the active picture; blanking intervals, file-format padding (some 10-bit formats pack samples into 32-bit words) and audio add a little more. Use it to check whether an interface or drive can sustain the data rate.

Common mistakes

Forgetting chroma samples and counting only luma.
Mixing bits and bytes when comparing with interface speeds.
Comparing uncompressed sizes with compressed recording bitrates.
Using MB and MiB interchangeably.

Keep in mind: Real file sizes depend on container formats, packing and compression. Check your recorder's or codec's specifications for exact storage needs.

Frequently asked questions

What does 4:2:0 mean?

Colour is stored at half the resolution horizontally and vertically, averaging 1.5 samples per pixel.

Why is 10-bit video larger than 8-bit?

Each sample uses 10 bits instead of 8, so data grows by 25 percent.

Why are camera files so much smaller?

Cameras compress video, removing redundancy between and within frames.

Does audio change the result much?

Uncompressed audio adds a few megabits per second, small compared with 4K video.

Sources: Recommendation ITU-R BT.2020-2 (10/2015), picture spatial characteristics, frame frequencies and sampling formats · IEC 80000-13 (binary prefixes such as MiB)

Checked 9 Nov 2026. Recommendation ITU-R BT.2020-2 (10/2015): Table 1 picture spatial characteristics (3840 × 2160, 7680 × 4320), Table 2 frame frequencies, Table 5 sampling formats 4:4:4, 4:2:2 and 4:2:0 and 10- or 12-bit coding; IEC 80000-13 binary prefixes (MiB).

Your numbers