libavif

libavif - Library for encoding and decoding .avif files

1,701

214

1,701

147

View on GitHub

Top Related Projects

libheif

1,895

libheif is an HEIF and AVIF file format decoder and encoder.

rav1e

3,823

The fastest and safest AV1 encoder.

vmaf

4,838

Perceptual video quality assessment based on multi-method fusion.

squoosh

22,675

Make images smaller using best-in-class codecs, right in the browser.

Quick Overview

The AOMediaCodec/libavif repository is a C library that provides an implementation of the AV1 Image File Format (AVIF), a modern image format based on the AV1 video codec. The library aims to provide a simple and efficient way to work with AVIF images, including encoding, decoding, and metadata manipulation.

Pros

Efficient Compression: AVIF offers significantly better compression compared to traditional image formats like JPEG, resulting in smaller file sizes without sacrificing image quality.
Wide Adoption: AVIF is supported by major web browsers and is gaining traction as a modern image format for the web and other applications.
Flexible Metadata: The AVIF format supports a wide range of metadata, including ICC color profiles, EXIF data, and more.
Cross-Platform Compatibility: The libavif library is designed to be cross-platform, with support for various operating systems and architectures.

Cons

Limited Browser Support: While AVIF is gaining traction, it is not yet universally supported by all web browsers, which can limit its adoption in some web-based applications.
Complexity: Implementing AVIF support can be more complex than working with simpler image formats like JPEG or PNG, especially for developers new to the format.
Performance Overhead: Encoding and decoding AVIF images can be more computationally intensive than other image formats, which may impact performance in some use cases.
Licensing Considerations: The AV1 codec, which is the basis for the AVIF format, has some licensing and patent considerations that developers should be aware of.

Code Examples

Here are a few examples of how to use the libavif library:

Decoding an AVIF Image:

#include <avif/avif.h>

avifImage* image = avifImageCreateEmpty();
avifDecoder* decoder = avifDecoderCreate();
avifResult result = avifDecoderReadImage(decoder, image);
if (result == AVIF_RESULT_OK) {
    // The image has been successfully decoded and can be accessed through the `image` object
}
avifImageDestroy(image);
avifDecoderDestroy(decoder);

Encoding an AVIF Image:

#include <avif/avif.h>

avifImage* image = avifImageCreate(width, height, 8, AVIF_PIXEL_FORMAT_YUV420);
// Set image data and metadata
avifEncoder* encoder = avifEncoderCreate();
encoder->speed = AVIF_SPEED_SLOWEST;
encoder->quality = 50;
avifResult result = avifEncoderWrite(encoder, image, &data, &size);
if (result == AVIF_RESULT_OK) {
    // The AVIF image data is available in the `data` buffer
}
avifImageDestroy(image);
avifEncoderDestroy(encoder);

Extracting Metadata from an AVIF Image:

#include <avif/avif.h>

avifImage* image = avifImageCreateEmpty();
avifDecoder* decoder = avifDecoderCreate();
avifResult result = avifDecoderReadImage(decoder, image);
if (result == AVIF_RESULT_OK) {
    // Access metadata properties of the image, such as:
    printf("Image width: %d\n", image->width);
    printf("Image height: %d\n", image->height);
    printf("Bit depth: %d\n", image->depth);
    printf("Color profile: %s\n", image->colorProfile.data);
}
avifImageDestroy(image);
avifDecoderDestroy(decoder);

Getting Started

To get started with the libavif library, follow these steps:

Clone the repository:

git clone https://github.com/AOMediaCodec/libavif.git

Build the library:

cd libavif
mkdir build
cd build
cmake ..
make

Install the library:

sudo make install

Competitor Comparisons

libheif

1,895

libheif is an HEIF and AVIF file format decoder and encoder.

Pros of libheif

Supports both HEIF and AVIF formats, offering more flexibility
More mature project with longer development history
Includes a simple API for encoding and decoding images

Cons of libheif

Larger codebase, potentially more complex to maintain
May have slower performance for AVIF-specific operations
Requires additional dependencies for full functionality

Code Comparison

libheif:

heif_context* ctx = heif_context_alloc();
heif_context_read_from_file(ctx, input_filename, nullptr);
heif_image_handle* handle;
heif_context_get_primary_image_handle(ctx, &handle);
heif_image* img;
heif_decode_image(handle, &img, heif_colorspace_RGB, heif_chroma_interleaved_RGB, nullptr);

libavif:

avifDecoder * decoder = avifDecoderCreate();
avifResult result = avifDecoderSetIOFile(decoder, input_filename);
result = avifDecoderParse(decoder);
avifImage * image = avifImageCreateEmpty();
result = avifDecoderNextImage(decoder);
avifRGBImage rgb;
avifRGBImageSetDefaults(&rgb, image);
avifImageYUVToRGB(image, &rgb);

Both libraries provide similar functionality for decoding images, but libavif focuses specifically on AVIF format, while libheif supports multiple formats. The APIs differ in structure and naming conventions, but the overall process is comparable.

rav1e

3,823

The fastest and safest AV1 encoder.

Pros of rav1e

Written in Rust, offering memory safety and performance benefits
Focused specifically on AV1 encoding, potentially leading to optimized performance
Active development with frequent updates and improvements

Cons of rav1e

More limited in scope compared to libavif's broader AVIF support
May have a steeper learning curve for developers not familiar with Rust
Potentially less mature ecosystem and tooling compared to libavif

Code Comparison

rav1e (Rust):

use rav1e::prelude::*;

let cfg = Config::default();
let mut ctx: Context<u8> = cfg.new_context().unwrap();
let frame = ctx.new_frame();
ctx.send_frame(frame).unwrap();
let packet = ctx.receive_packet().unwrap();

libavif (C):

#include "avif/avif.h"

avifEncoder * encoder = avifEncoderCreate();
avifRWData avifOutput = AVIF_DATA_EMPTY;
avifImage * image = avifImageCreate(width, height, depth, format);
avifResult result = avifEncoderWrite(encoder, image, &avifOutput);

Both libraries provide encoding capabilities for AV1-based formats, but rav1e focuses on AV1 encoding while libavif offers broader AVIF support. rav1e's Rust implementation may provide better memory safety, while libavif's C implementation might be more widely compatible with existing codebases.

vmaf

4,838

Perceptual video quality assessment based on multi-method fusion.

Pros of VMAF

Focused on video quality assessment, providing a perceptual metric
Includes pre-trained models for different use cases
Widely adopted in the streaming industry, particularly by Netflix

Cons of VMAF

Limited to quality assessment, not an image/video codec like libavif
Requires more computational resources for analysis
May not be as suitable for real-time applications

Code Comparison

VMAF (Python):

import vmaf

model = vmaf.VmafModel()
score = model.compute_score(reference_video, distorted_video)

libavif (C):

avifImage *image = avifImageCreate(width, height, depth, format);
avifEncoder *encoder = avifEncoderCreate();
avifResult result = avifEncoderWrite(encoder, image, &raw);

Key Differences

Purpose: VMAF is for video quality assessment, while libavif is an AV1 Image File Format library
Language: VMAF is primarily Python-based, libavif is C-based
Functionality: VMAF analyzes video quality, libavif encodes/decodes AVIF images
Use cases: VMAF is used for streaming optimization, libavif for next-gen image compression

Both projects serve different purposes in the multimedia ecosystem, with VMAF focusing on quality metrics and libavif on image format implementation.

squoosh

22,675

Make images smaller using best-in-class codecs, right in the browser.

Pros of Squoosh

User-friendly web interface for image compression and conversion
Supports multiple image formats and compression algorithms
Provides real-time preview and comparison of compressed images

Cons of Squoosh

Limited to browser-based usage, not suitable for server-side or command-line operations
Fewer advanced options and customization compared to libavif
Primarily focused on image compression, while libavif specializes in AVIF format

Code Comparison

libavif (C):

avifResult result = avifEncoderWrite(encoder, image, &raw);
if (result != AVIF_RESULT_OK) {
    fprintf(stderr, "Failed to encode: %s\n", avifResultToString(result));
    return 1;
}

Squoosh (JavaScript):

const compressedImage = await imagePool.ingestImage(imageFile);
await compressedImage.encode({
  webp: {},
  mozjpeg: {},
});

The code snippets demonstrate the core functionality of each project. libavif focuses on AVIF encoding, while Squoosh provides a more general-purpose image compression interface supporting multiple formats.

Convert designs to code with AI

Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.

Try Visual Copilot

README

libavif

This library aims to be a friendly, portable C implementation of the AV1 Image File Format, as described here:

https://aomediacodec.github.io/av1-avif/

It can encode and decode all AV1 supported YUV formats and bit depths (with alpha). In addition to the library, encoder and decoder command line tools are also provided (avifenc and avifdec).

It is recommended that you check out/use tagged releases instead of just using the main branch. We will regularly create new versions as bug fixes and features are added.

Command line tool usage

avifenc -q 75 input.[jpg|png|y4m] output.avif
avifdec output.avif decoded.png

See avifenc --help for all options.

API usage

Please see the examples in the "examples" directory. If you're already building libavif, enable the CMake option AVIF_BUILD_EXAMPLES in order to build and run the examples too.

Installation

libavif is a package in most major OSs.

Windows

vcpkg install libavif

You can also download the official windows binaries on the release page.

macOS

Homebrew:

brew install libavif

MacPorts:

sudo port install libavif

Linux

Debian-based distributions:

sudo apt install libavif-dev

Red Hat-based distributions:

sudo yum -y install libavif

MinGW

For the "default" MSYS2 UCRT64 environment:

pacman -S mingw-w64-ucrt-x86_64-libavif

Build Notes

Building libavif requires CMake. See Build Command Lines below for example command lines.

Controlling Dependencies

CMake flags like AVIF_CODEC_AOM, AVIF_LIBYUV, etc. allow enabling or disabling dependencies. They can take three possible values:

OFF: the dependency is disabled.
SYSTEM: the dependency is expected to be installed on the system.
LOCAL: the dependency is built locally. In most cases, CMake can automatically download and build it. For some dependencies, you need to run the associated script in the ext/ subdirectory yourself. In cases where CMake handles downloading the dependency, you can still call the script in ext/ if you want to use a different version of the dependency (e.g. by modifying the script) or make custom code changes to it. If a directory with the dependency exists in the ext/ directory, CMake will use it instead of downloading a new copy.

Codec Dependencies

No AV1 codecs are enabled by default. You should enable at least one of them by setting any of the following CMake options to LOCAL or SYSTEM, depending on whether you want to use a locally built or a system installed version (e.g. -DAVIF_CODEC_AOM=LOCAL):

AVIF_CODEC_AOM for libaom (encoder and decoder)
AVIF_CODEC_DAV1D for dav1d (decoder)
AVIF_CODEC_LIBGAV1 for libgav1 (decoder)
AVIF_CODEC_RAV1E for rav1e (encoder)
AVIF_CODEC_SVT for SVT-AV1 (encoder)

When set to SYSTEM, these libraries (in their C API form) must be externally available (discoverable via CMake's FIND_LIBRARY) to use them, or if libavif is a child CMake project, the appropriate CMake target must already exist by the time libavif's CMake scripts are executed.

Libyuv Dependency

Libyuv is an optional but strongly recommended dependency that speeds up color space conversions. It's enabled by default with a value of SYSTEM, so it's expected to be installed on the system. It can either be built locally instead by using -DAVIF_LIBYUV=LOCAL or disabled with -DAVIF_LIBYUV=OFF.

Tests

A few tests written in C can be built by enabling the AVIF_BUILD_TESTS CMake option.

The remaining tests require GoogleTest, and can be built by enabling AVIF_BUILD_TESTS and setting AVIF_GTEST to SYSTEM or LOCAL.

Additionally, fuzzing tests require fuzztest, see also fuzzing test instructions in ext/oss-fuzz/README.md.

Code coverage is available by enabling AVIF_ENABLE_COVERAGE then building the avif_coverage target, e.g. make avif_coverage -j. It requires compiling with clang (-DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++) and LLVM must be installed on the system.

Build Command Lines {#build-command-lines}

The following instructions can be used to build the libavif library and the avifenc and avifdec tools.

Build using installed dependencies

To link against the already installed aom, libjpeg, libpng and libyuv dependency libraries (recommended):

git clone -b v1.2.1 https://github.com/AOMediaCodec/libavif.git
cmake -S libavif -B libavif/build -DAVIF_CODEC_AOM=SYSTEM -DAVIF_BUILD_APPS=ON
cmake --build libavif/build --config Release --parallel

Build everything from scratch

For development and debugging purposes:

git clone -b v1.2.1 https://github.com/AOMediaCodec/libavif.git
cmake -S libavif -B libavif/build -DCMAKE_BUILD_TYPE=Debug -DBUILD_SHARED_LIBS=OFF -DAVIF_CODEC_AOM=LOCAL -DAVIF_LIBYUV=LOCAL -DAVIF_LIBSHARPYUV=LOCAL -DAVIF_JPEG=LOCAL -DAVIF_ZLIBPNG=LOCAL -DAVIF_BUILD_APPS=ON
cmake --build libavif/build --config Debug --parallel

Prebuilt Binaries (Windows)

Statically-linked avifenc.exe and avifdec.exe can be downloaded from the Releases page.

Development Notes

Please check the wiki for extra resources on libavif, such as the Release Checklist.

The libavif library is written in C99. Most of the tests are written in C++14.

Formatting

Use clang-format to format the sources from the top-level folder (clang-format-16 preferred):

clang-format -style=file -i \
  apps/*.c apps/*/*.c apps/*/*.cc apps/*/*.h examples/*.c \
  include/avif/*.h src/*.c src/*.cc \
  tests/*.c tests/*/*.cc tests/*/*.h

Use cmake-format to format the CMakeLists.txt files from the top-level folder:

cmake-format -i \
  CMakeLists.txt \
  tests/CMakeLists.txt \
  cmake/Modules/*.cmake \
  contrib/CMakeLists.txt \
  contrib/gdk-pixbuf/CMakeLists.txt \
  android_jni/avifandroidjni/src/main/jni/CMakeLists.txt

License

Released under the BSD License.

Copyright 2019 Joe Drago. All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:

1. Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.

2. Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

Top Related Projects

Convert designs to code with AI

Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.

Try Visual Copilot