Bit depth

What the number counts

Bit depth is normally quoted per channel. An 8-bit image has 256 levels of red, 256 of green and 256 of blue, giving about 16.7 million colours in total — which is where the marketing phrase "24-bit colour" comes from, three channels at eight bits each.

A 16-bit image has 65,536 levels per channel. It does not look better on a screen showing the same photograph; what it has is far more room between adjacent levels.

Why the headroom matters while editing

Every tonal adjustment redistributes values. Lifting shadows stretches a narrow range of dark values across a wider range of output values, and in an 8-bit file there may be only a dozen distinct levels down there to stretch. The result is gaps, and gaps in a gradient are banding.

In a 16-bit file the same adjustment has hundreds of levels to work with and the gradient survives intact. This is the practical case for editing at higher bit depth even when the output will be 8-bit: you are protecting against the damage done between opening the file and saving it, not against the final save.

Convert to 8-bit at export, once, after the editing is done.

Where it shows up in practice

Raw camera files are typically 12 or 14 bits per channel, which is why a raw file tolerates exposure correction that would visibly break a JPEG from the same camera.

JPEG is 8-bit only. PNG supports 8 and 16. WebP is 8-bit. AVIF supports 10 and 12 bits, which is part of why it handles gradients so much better — it has both a better codec and more levels to spend.

Screens are catching up unevenly: most panels are 8-bit, some are 10-bit, and content graded for high dynamic range depends on the extra depth to avoid visible steps in bright gradients.