RAW to ACES Utility

Table of Contents

1. Introduction
2. Package Contents
3. Prerequisites
4. Installation
5. Usage
6. Known Issues
7. Installation for Redhat
8. License

Introduction

The RAW to ACES Utility or rawtoaces, is a software package that converts digital camera RAW files to ACES container files containing image data encoded according to the Academy Color Encoding Specification (ACES) as specified in SMPTE 2065-1. This is accomplished through one of two methods.

1. CameraRAW RGB data (generated by libraw) is converted to ACES by calculating an Input Device Transform (IDT) based on the camera's sensitivity and a light source.

2. CameraRAW RGB data (generated by libraw) is converted to ACES by calculating an RGB to XYZ matrix using information included in libraw and metadata found in the RAW file.

The output image complies with the ACES Container specification (SMPTE S2065-4).

Package Contents

The source code contains the following:

• cmake/ - CMake modules for locating dependencies (e.g., libraw)
• config/ - CMake configuration files
• data/ - Data files containing camera sensitivity, light source, color matching function and 190 training patch data
• lib/ - IDT and math libraries
• src/ - AcesRender wrapper library and C++ header file containing rawtoaces usage information
• test/ - Sample testing materials such as a ".NEF" RAW image and a camera spectral sensitivity data file
• main.cpp - C++ source code file for call routines to process images

Prerequisites

To build rawtoaces you would need to sutisfy these dependencies:

Library	Min Version	Purpose	Link to installation instruction
cmake	3.10		CMake download
ceres	1.12.0	Ceres Solver is an open source library for solving Non-linear Least Squares problems with bounds constraints and unconstrained optimization problems. It processes non-linear regression for rawtoaces.	Ceres Solver installation
imath	3.1.8	Provides the half data type used for representing 16-bit floating-point values. It's used by rawtoaces for storing high dynamic range (HDR) data in a compact format.	Imath installation
libraw	0.19.4	LibRaw is a library that processes RAW files from digital cameras. It handles image pre-processing for rawtoaces.	LibRaw download
boost	1.76.0	Boost has multiple C++ libraries that support tasks related to linear algebra, multithreading, image processing, unit testing, etc. It handles data loading and unit testing for rawtoaces.	Boost download
aces_container	latest	ACES Container is the reference implementation for a file writer intended to be used with the Academy Color Encoding System (ACES). rawtoaces relies on it to produce images that comply with the ACES container specification (SMPTE S2065-4).	ACES Container installation

MacOS

Install homebrew if not already installed

$ /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"

$ brew install cmake ceres-solver imath libraw boost aces_container  

Linux

$ sudo apt-get -f cmake
$ build_scripts/install_aces_container.bash
$ sudo apt-get -f install \
    libimath-dev \
    libboost-dev libboost-filesystem-dev \
    libboost-test-dev \
    libraw-dev libceres-dev

RedHat

$ sudo yum install cmake
$ build_scripts/install_aces_container.bash
$ sudo yum install eigen3-devel ceres-solver-devel LibRaw-devel boost-devel imath

Windows

Install cmake first via installer or with Chocolatey:

$ choco install cmake

$ build_scripts/install_aces_container.bash
$ vcpkg install \
    libraw:x64-windows \
    ceres:x64-windows \
    imath:x64-windows \
    boost-system:x64-windows \
    boost-foreach:x64-windows \
    boost-filesystem:x64-windows \
    boost-test:x64-windows \
    boost-property-tree:x64-windows

Build and Install

From Source

From the root source directory:

$ git clone https://github.com/ampas/rawtoaces
$ cd rawtoaces
$ cmake -S . -B build
$ cmake --build build
$ sudo cmake --install build # Optional if you want it to be accessible system wide

The default process will install librawtoaces_idt_${rawtoaces_version}.dylib and librawtoaces_util_${rawtoaces_version}.dylib to /usr/local/lib, a few header files to /usr/local/include/rawtoaces/include and a number of data files into /usr/local/include/rawtoaces/data.

Docker

Assuming you have Docker installed, installing and running rawtoaces is relatively straightforward except for the compilation of ceres-solver that requires a significant amount of memory, well over the 2Go that Docker allocates by default. Thus you will need to increase the memory in Docker preferences: Preferences --> Resources --> Advanced, 8Go should be enough.

• From the root source directory, build the container:

  sh $ docker build -f "Dockerfile" -t rawtoaces:latest "."

• Then to run it from a directory with images:

  sh $ docker run -it --rm -v $PWD:/tmp -w /tmp rawtoaces:latest rawtoaces IMG_1234.CR2

Usage

Overview

rawtoaces uses one of three methods to convert RAW image files to ACES.

1. Camera spectral sensitivities and illuminant spectral power distributions
2. Camera file metadata
3. Camera data included in the libraw software

The preferred, and most accurate, method of converting RAW image files to ACES is to use camera spectral sensitivities and illuminant spectral power distributions, if available. If spectral sensitivity data is available for the camera, rawtoaces uses the method described in Academy document P-2013-001 (.pdf download).

While preferred, camera spectral sensitivity data is not commonly known to general users. When that is the case, rawtoaces can use either the metadata embedded in the camera file or camera data included in libraw to approximate a conversion to ACES.

Help message

A help message with a description of all command line options can be obtained by typing the following command:

$ rawtoaces --help
rawtoaces - convert RAW digital camera files to ACES

Usage:
  rawtoaces file ...
  rawtoaces [options] file
  rawtoaces --help
  rawtoaces --version

IDT options:
  --help                  Show this screen
  --version               Show version
  --wb-method [0-4]       White balance factor calculation method
                            0=white balance using file metadata 
                            1=white balance using user specified illuminant [str] 
                            2=Average the whole image for white balance
                            3=Average a grey box for white balance <x y w h>
                            4=Use custom white balance  <r g b g>
                            (default = 0)
  --mat-method [0-2]      IDT matrix calculation method
                            0=Calculate matrix from camera spec sens
                            1=Use file metadata color matrix
                            2=Use adobe coeffs included in libraw
                            (default = 0)
    --headroom float        Set highlight headroom factor (default = 6.0)
    --cameras               Show a list of supported cameras/models by LibRaw
    --valid-illums          Show a list of illuminants
    --valid-cameras         Show a list of cameras/models with available 
                          spectral sensitivity datasets

Raw conversion options:
  -c float                Set adjust maximum threshold (default = 0.75)
  -C <r b>                Correct chromatic aberration
  -P <file>               Fix the dead pixels listed in this file
  -K <file>               Subtract dark frame (16-bit raw PGM)
  -k <num>                Set the darkness level
  -S <num>                Set the saturation level
  -n <num>                Set threshold for wavelet denoising
  -H [0-9]                Highlight mode (0=clip, 1=unclip, 2=blend, 3+=rebuild) (default = 0)
  -t [0-7]                Flip image (0=none, 3=180, 5=90CCW, 6=90CW)
  -j                      Don't stretch or rotate raw pixels
  -W                      Don't automatically brighten the image
  -b <num>                Adjust brightness (default = 1.0)
  -q [0-3]                Set the interpolation quality
  -h                      Half-size color image (twice as fast as "-q 0")
  -f                      Interpolate RGGB as four colors
  -m <num>                Apply a 3x3 median filter to R-G and B-G
  -s [0..N-1]             Select one raw image from input file
  -G                      Use green_matching() filter
  -B <x y w h>            Use cropbox

Benchmarking options:
  -v                      Verbose: print progress messages (repeated -v will add verbosity)
  -F                      Use FILE I/O instead of streambuf API
  -d                      Detailed timing report
  -E                      Use mmap()-ed buffer instead of plain FILE I/O

RAW conversion options

In most cases the default values for all "RAW conversion options" should be sufficient. Please see the help menu for details of the RAW conversion options.

Conversion using spectral sensitivities

If spectral sensitivity data for your camera is included with rawtoaces then the following command will convert your RAW file to ACES using that information.

$ rawtoaces input.raw

This command is equivalent to :

$ rawtoaces --wb-method 0 --mat-method 0 input.raw

To process mutiple raw files, you can try:

$ rawtoaces input1.raw input2.raw

To batch-process raw files in a directory, you can try:

$ rawtoaces input_dir

To batch-process raw files from multiple directories, you can try:

$ rawtoaces input_dir1 input_dir2

This is the preferred method as camera white balance gain factors and the RGB to ACES conversion matrix will be calculated using the spectral sensitivity data from your camera. This provides the most accurate conversion to ACES.

By default, rawtoaces will determine the adopted white by finding the set of white balance gain factors calculated from spectral sensitivities closest to the "As Shot" (aka Camera Multiplier) white balance gain factors included in the RAW file metadata. This default behavior can be overridden by including the desired adopted white name after the white balance method. The following example will use the white balance gain factors calculated from spectral sensitivities for D60.

$ rawtoaces --wb-method 1 D60 --mat-method 0 input.raw

You can use the environment varilable of AMPAS_DATA_PATH to specify the repository for your own datasets. If you have spectral sensitivity data for your camera but it is not included with rawtoaces you may place that data in /usr/local/include/rawtoaces/data/camera or place the data in the folder pointed by AMPAS_DATA_PATH.

JSON Schema for Spectral Datasets

The schema takes its roots in IES TM-27-14 but implements support for multiple spectral datasets while adopting JSON over XML for the simplicity of its grammar.

rawtoaces expects the spectral dataset to have the following shape: (380, 780, 5), i.e. starting from 380nm and ending at 780nm with a 5nm interval/bin size.

The attributes are defined as follows:

• header
  • schema_version
    • description: Required, schema version of the current document.
    • type: "string"
  • catalog_number
    • description: Optional,
    • type: ["string", "null"]
  • comments
    • description: Optional, additional information for the spectral dataset.
    • type: ["string", "null"]
  • description
    • description: Required, description of the spectral dataset.
    • type: "string"
  • document_creation_date
    • description: Required, document creation date expressed as per RFC 3339 - Date and Time on the Internet: Timestamps, e.g. 2017-01-01T12:00:00Z.
    • type: "string"
    • format: "date-time"
  • document_creator
    • description: Required, creator of the document, e.g. company, individual, laboratory, etc.
    • type: "string"
  • laboratory
    • description: Optional, laboratory or company that performed the measurements.
    • type: ["string", "null"]
  • license
    • description: Required, usage license of the document, e.g. "CC-BY-NC-ND".
    • type: ["string", "null"]
  • manufacturer
    • description: Optional, manufacturer of the device being tested.
    • type: ["string", "null"]
  • measurement_equipment
    • description: Optional, measurement equipment used to test the device.
    • type: ["string", "null"]
  • model
    • description: Optional, model of the device being tested.
    • **ty