A fast and lightweight GIF encoder that can create GIF animations and images. Summary of the main features:
To get started, we suggest that you have a look at our code examples. examples/cgif_example_video.c
is an example that creates a GIF animation. examples/cgif_example.c
is an example for a static GIF image.
To get an overview of the API, we recommend having a look at our wiki (https://github.com/dloebl/cgif/wiki/General-API) where types and functions are described. The corresponding implementations can be found in src/cgif.c
and src/cgif_raw.c
. Here the most important types and functions:
// These are the four struct types that contain all GIF data and parameters:
typedef CGIF_Config // global cofinguration parameters of the GIF
typedef CGIF_FrameConfig // local configuration parameters for a frame
typedef CGIF // struct for the full GIF
typedef CGIF_Frame // struct for a single frame
// The user needs only these three functions to create a GIF image:
CGIF* cgif_newgif (CGIF_Config* pConfig); // creates a new GIF
int cgif_addframe (CGIF* pGIF, CGIF_FrameConfig* pConfig); // adds a frame to an existing GIF
int cgif_close (CGIF* pGIF); // close the created file and free memory
With our encoder you can create animated or static GIFs, you can or cannot use certain optimizations, and so on. You can switch between all these different options easily using the two attributes attrFlags
and genFlags
in the configurations CGIF_Config
and CGIF_FrameConfig
. These attributes are of type uint32_t
and bundle yes/no-options with a bit-wise logic. So far only a few of the 32 bits are used leaving space to include further functionalities ensuring backward compatibility. We provide the following flag settings which can be combined by bit-wise or-operations:
CGIF_ATTR_IS_ANIMATED // make an animated GIF (default is non-animated GIF)
CGIF_ATTR_NO_GLOBAL_TABLE // disable global color table (global color table is default)
CGIF_ATTR_HAS_TRANSPARENCY // first entry in color table contains transparency (alpha channel)
CGIF_ATTR_NO_LOOP // run GIF animation only one time. numLoops is ignored (no repetitions)
CGIF_GEN_KEEP_IDENT_FRAMES // keep frames that are identical to previous frame (default is to drop them)
CGIF_FRAME_ATTR_USE_LOCAL_TABLE // use a local color table for a frame (not used by default)
CGIF_FRAME_ATTR_HAS_ALPHA // frame contains alpha channel (index set via transIndex field)
CGIF_FRAME_ATTR_HAS_SET_TRANS // transparency setting provided by user (transIndex field)
CGIF_FRAME_ATTR_INTERLACED // encode frame interlaced
CGIF_FRAME_GEN_USE_TRANSPARENCY // use transparency optimization (size optimization)
CGIF_FRAME_GEN_USE_DIFF_WINDOW // do encoding just for the sub-window that changed (size optimization)
If you didn't understand the point of attrFlags
and genFlags
and the flags, please don't worry. The example files examples/cgif_example.c
and examples/cgif_example_video.c
are all you need to get started and the used default settings for attrFlags
and genFlags
cover most cases quite well.
An example can be compiled and tested simply by:
$ c99 -o cgif_example -Iinc examples/cgif_example_video.c src/cgif.c src/cgif_raw.c
$ ./cgif_example
In the folder tests
, we provide several testing routines that you can run via the script tests/performtests.sh
. To perform the tests you need to install the programs ImageMagick, gifsicle and tcc (tiny c compiler).
With the provided tests you can validate that the encoder still generates correct GIF files after making changes on the encoder itself.
The GIF format employs the Lempel-Ziv-Welch (LZW) algorithm for image compression. If you are interested in details of the GIF format, please have a look at the official GIF documentation (https://www.w3.org/Graphics/GIF/spec-gif89a.txt).
Releases of cgif follow the semantic versioning scheme as described here: semver.org
The following additional guarantees are provided:
There is a Visual Studio Code debug configuration with launch targets for the two examples. You need to install the C/C++ extension and a LLDB extension (debugger) to debug cgif.
Licensed under the MIT license (permissive).
For more details please see LICENSE