Crate gamut_color 

Color primitives for the gamut codecs: pixel formats, bit depths, chroma subsampling, the CICP code points carried in nclx / AV1 sequence headers, and planar buffers.

The M0 AVIF encoder uses only a narrow slice — 8-bit RGB in, mapped to identity (mc = 0) 4:4:4 planes. The enums here intentionally model the wider spec surface (more formats, bit depths, subsamplings, and CICP code points) so later milestones (M2 pixel formats, M4 HDR; see gamut-avif/STATUS.md) extend without reshaping the types.

On top of that metadata layer, the transfer, oklab, matrix, gamut_map, and profile modules add f64 colour science — encoder-exact EOTFs, OKLab transforms with per-gamut matrices (derived from chromaticities via Bradford adaptation), gamut clamping, and source-profile bundles over the CICP axes. This math is Tier-1 (correctness only): it uses std f64, so it is not bit-reproducible across platforms — see references/color/README.md.

Re-exports

pub use cicp::ColorRange;

pub use cicp::ColourPrimaries;

pub use cicp::MatrixCoefficients;

pub use cicp::TransferCharacteristics;

Modules

cicp

CICP (ITU-T H.273 / ISO/IEC 23091-2) code points shared by the AVIF colr (nclx) box and the AV1 sequence header color_config (AV1 §5.5.2 / §6.4.2).

gamut_map

sRGB gamut membership and a hue-preserving soft gamut clamp.

matrix

Derivation of RGB→XYZ and OKLab M1 matrices from CIE 1931 chromaticities, with Bradford chromatic adaptation.

oklab

OKLab transforms and the per-gamut linear-RGB → LMS (M1) matrices.

profile

Source-profile bundle: ties a colour-science Gamut together with its encoder-exact transfer (and tone map) and projects them back onto gamut’s independent CICP axes.

transfer

Transfer functions (EOTF / inverse-EOTF) in f64, covering the gamuts the gamut ecosystem encodes through.

Structs

Planar8

Three full-resolution (4:4:4) 8-bit planes, each width * height samples, row-major.

Yuv420

A BT.601 YCbCr image in 4:2:0 layout: a width × height luma plane and two chroma_width × chroma_height chroma planes, all row-major 8-bit, where the chroma dimensions are ceil(width / 2) and ceil(height / 2). The signal range is a property of the conversion (Bt601Range), not of the stored planes.

Enums

BitDepth

Bits per sample of a coded plane.

Bt601Range

The signal range of a BT.601 YCbCr encoding. See the module docs for which to use.

ChromaSubsampling

Chroma subsampling of the coded planes (AV1 subsampling_x / subsampling_y, §5.5.2).

Functions

clip_pixel

Saturates a computed integer sample to the unsigned bit_depth-bit pixel range 0..=(1 << bit_depth) - 1 (AV1 Clip1, §3) — the high-bit-depth companion to clip_pixel8.

clip_pixel8

Saturates a computed integer sample to the unsigned 8-bit pixel range 0..=255.

rgb_to_ycbcr

Converts one RGB triple to BT.601 YCbCr in the given Bt601Range (each component 0..=255).

ycbcr_to_rgb

Converts one BT.601 YCbCr triple in the given Bt601Range back to RGB (each 0..=255).