gamut-color

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

Goals

Part of the gamut workspace, this crate exists to:

• Centralize color management. One place defines [BitDepth], [ChromaSubsampling], the CICP enums ([ColourPrimaries], [TransferCharacteristics], [MatrixCoefficients], [ColorRange]), and the [Planar8] buffer — so a color bug is fixed once, not re-implemented per format. (The interleaved pixel layouts are the typed Pixel markers in gamut-core.)
• Model the full spec surface up front. The M0 AVIF path uses only a narrow slice (8-bit RGB in, mapped to identity mc = 0 4:4:4 planes), but the enums intentionally cover the wider range of 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.
• Match the spec code points exactly. CICP values mirror the H.273 / AV1 sequence-header code points so they round-trip through av1C/colr and AV1 headers unchanged.
• Stay memory-safe. #![forbid(unsafe_code)].

Usage

use gamut_color::Planar8;

// 8-bit interleaved RGB -> identity (mc = 0) 4:4:4 planes (Y=G, U=B, V=R).
let width = 2;
let height = 2;
let rgb: Vec<u8> = vec![0; width * height * 3];
let planes = Planar8::from_rgb8_identity(&rgb, width as u32, height as u32).expect("valid input");
assert_eq!(planes.width(), 2);
let _y = planes.plane(0); // luma plane

Status

M0 exercises the 8-bit RGB ↔ identity 4:4:4 conversion ([Planar8::from_rgb8_identity] / [Planar8::to_rgb8_identity]) plus the CICP tables the AVIF colr box needs. The remaining formats, bit depths, and subsamplings are modeled in the type system but not yet wired into an encode path; they land with the milestones tracked in gamut-avif/STATUS.md.

License

Licensed under either of MIT or Apache-2.0 at your option.