Background

libultrahdr is an image compression library that uses gain map technology to store and distribute HDR images. Conceptually on the encoding side, the library accepts SDR and HDR rendition of an image and from these a Gain Map (quotient between the two renditions) is computed. The library then uses backward compatible means to store the base image (SDR), gain map image and some associated metadata. Legacy readers that do not support handling the gain map image and/or metadata, will display the base image. Readers that support the format combine the base image with the gain map and render a high dynamic range image on compatible displays.

For additional information about libultrahdr, see android hdr-image-format guide.

Building libultrahdr

Requirements

• CMake v3.13 or later
• C++ compiler, supporting at least C++17.
• libultrahdr uses jpeg compression format to store sdr image and gainmap quotient. So, libjpeg or any other jpeg codec that is ABI and API compatible with libjpeg.

The library offers a way to skip installing libjpeg by passing UHDR_BUILD_DEPS=1 at the time of configure. That is, cmake -DUHDR_BUILD_DEPS=1 will clone jpeg codec from link and include it in the build process. This is however not recommended.

If jpeg is included in the build process then to build jpeg with simd extensions, - C compiler - NASM or Yasm are needed. * If using NASM, 2.13 or later is required. * If using Yasm, 1.2.0 or later is required.

Build Procedure

To build libultrahdr, examples, unit tests:

Un*x (including Linux, Mac)

mkdir build_directory
cd build_directory
cmake -G "Unix Makefiles" -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DUHDR_BUILD_TESTS=1 ../
make
ctest
make install

This will generate the following files under build_directory:

libuhdr.so or libuhdr.dylib
ultrahdr shared library

libuhdr.pc
ultrahdr pkg-config file

ultrahdr_app
Statically linked sample application demonstrating ultrahdr API usage

ultrahdr_unit_test
Unit tests

make install will install libuhdr.so, ultrahdr_api.h, libuhdr.pc for system-wide usage and make uninstall will remove the same.

NOTE: you may need to run ldconfig after install/uninstall

MinGW

NOTE: This assumes that you are building on a Windows machine using the MSYS environment.

mkdir build_directory
cd build_directory
cmake -G "MSYS Makefiles" -DUHDR_BUILD_TESTS=1 ../
cmake --build ./
ctest

mkdir build_directory
cd build_directory
cmake -G "MinGW Makefiles" -DUHDR_BUILD_TESTS=1 ../
cmake --build ./
ctest

This will generate the following files under build_directory:

libuhdr.dll
ultrahdr shared library

ultrahdr_app.exe
Sample application demonstrating ultrahdr API

ultrahdr_unit_test.exe
Unit tests

Visual C++ (IDE)

mkdir build_directory
cd build_directory
cmake -G "Visual Studio 16 2019" -DUHDR_BUILD_DEPS=1 -DUHDR_BUILD_TESTS=1 ../
cmake --build ./ --config=Release
ctest -C Release

This will generate the following files under build_directory/Release:

ultrahdr_app.exe
Sample application demonstrating ultrahdr API

ultrahdr_unit_test.exe
Unit tests

Visual C++ (Command line)

mkdir build_directory
cd build_directory
cmake -G "NMake Makefiles" -DUHDR_BUILD_DEPS=1 -DUHDR_BUILD_TESTS=1 ../
cmake --build ./
ctest

This will generate the following files under build_directory:

ultrahdr_app.exe
Sample application demonstrating ultrahdr API

ultrahdr_unit_test.exe
Unit tests

NOTE: To not build unit tests, skip passing -DUHDR_BUILD_TESTS=1

Building Benchmark

To build benchmarks, pass -DUHDR_BUILD_BENCHMARK=1 to cmake configure command and build.

This will additionally generate,

ultrahdr_bm
Benchmark tests

Building Fuzzers

Refer to README.md for complete instructions.

Using libultrahdr

A detailed description of libultrahdr encode and decode api is included in ultrahdr_api.h and for sample usage refer demo app.

libultrahdr includes two classes of APIs, one to compress and the other to decompress HDR images:

Encoding api outline:

[^1]: Tonemap hdr to sdr. Compute gain map from hdr and sdr. Compress sdr and gainmap at quality configured. Add exif if provided. Combine sdr compressed, gainmap in multi picture format with gainmap metadata. [^2]: Compute gain map from hdr and sdr. Compress sdr and gainmap at quality configured. Add exif if provided. Combine sdr compressed, gainmap in multi picture format with gainmap metadata. [^3]: Compute gain map from hdr and raw sdr. Compress gainmap. Combine sdr compressed, gainmap in multi picture format with gainmap metadata. [^4]: Decode compressed sdr input. Compute gain map from hdr and decoded sdr. Compress gainmap. Combine sdr compressed, gainmap in multi picture format with gainmap metadata. [^5]: Combine sdr compressed, gainmap in multi picture format with gainmap metadata.

Decoding api outline:

Configure display device characteristics (display transfer characteristics, max display boost) for optimal usage.