Colour Depth & Raw Image Size

Bits per pixel = channels × bits per channel: RGB at 8 bits is 24 bpp, RGBA at 16 bits is 64. A pixel can take 2^bpp distinct values (16,777,216 at 24 bpp; 2^64 is shown in scientific form), and each channel has 2^bits levels — the tonal steps of the image, which is a different question from what a display can reproduce.

Uncompressed bytes = width × height × bpp ÷ 8. Some formats pad each row to an alignment (BMP to 4 bytes), and the page adds that when you give the alignment. For a Y′CbCr image, 4:2:2 and 4:2:0 subsampling keep luma at full resolution and thin the two chroma channels to an average of 2/3 or 1/2 of the 4:4:4 bits per pixel — 12 bpp instead of 24 for 4:2:0 — so the same rule applies to three-channel images only.

With a frame rate the raw data rate is frame bytes × 8 × fps, and a duration gives the total: 1080p 4:2:0 at 30 fps is about 746 Mbit/s before compression. A codec’s output size is set by its bit rate, not by this figure. Headers, TIFF tags, sensor-raw packing and lossless compression all change the file on disk. Nothing leaves the browser; the same four anonymous usage counts as the rest of the site apply.