Crate rawshift_image 

rawshift-image

Still-image support for rawshift: a high-performance RAW image processing library with support for multiple camera formats and a full processing pipeline, plus standard compressed-format decode/encode.

This crate is normally consumed through the rawshift facade crate (enable its image feature). Depend on rawshift-image directly when you need per-format Cargo feature control.

Supported Formats

RAW Formats (full pipeline)

• Sony ARW (v1–v5)
• Adobe DNG (v1.7, including Apple ProRAW)
• Canon CR2
• Canon CR3 (metadata + format detection; pixel decode pending CRX codec)
• Nikon NEF
• Fujifilm RAF

Standard Formats (direct RGB decode)

• GIF, JPEG, PNG, WebP, JPEG XL, TIFF
• SVG (requires svg feature)
• AVIF decode + encode (requires avif feature)
• HEIC decode (requires heic feature; via libheif)
• APV (detection only; no Rust decoder exists yet)

Quick Start

// Requires features = ["experimental"]
use rawshift_image::formats::RawFile;
use std::fs::File;

let file = File::open("image.arw").expect("Failed to open file");
let raw = RawFile::open(file).expect("Failed to parse RAW file");
let metadata = raw.metadata();
println!(
    "Camera: {} {}",
    metadata.camera.make,
    metadata.camera.model
);

Processing Pipeline

Raw images go through these steps:

1. Format decoding (ARW, DNG, CR2, NEF, RAF)
2. Black level subtraction
3. White balance
4. Demosaicing (AMaZE, RCD, LMMSE, Markesteijn)
5. Color matrix application
6. Tone mapping / gamma

Feature Flags

Cargo features are organised in five tiers, high-level to low-level. Each tier is defined in terms of the tier below; only tier-4 features (and RAW tier-3 features) pull in an external crate.

1. Bundles — default, full, experimental, raw-stabilizing, raw-incomplete.
2. Formats — jpeg, png, webp, jxl, avif, dng, gif, tiff, heic, svg, arw, cr2, cr3, crw, nef, raf (decode + encode for that format).
3. Directions — jpeg-decode, jpeg-encode, arw-decode, … For compressed formats a direction feature aliases the default implementation; RAW formats have a single in-repo implementation.
4. Implementations — compressed formats only, named format-direction-impl (e.g. jpeg-decode-zune). Multiple may be enabled at once; the active backend is chosen via formats::DecodeOptions and formats::export::EncodeOptions.
5. Infrastructure — tiff-parser, serde, heic-vendored.

See the “Feature Flags” section of the README for the full hierarchy.

Modules

core

Core types and traits for image processing.

data

Data constants for image processing.

error

Error types for RAW image processing.

formats

RAW format decoders.

prelude

Convenience re-exports of the most commonly used types and functions.

processing

Pixel processing and color space conversion primitives

tiff

Low-level TIFF engine.

transforms

Image transformation and processing pipeline.