MiniVideo framework
MiniVideo is a multimedia framework developed from scratch in C/C++, bundled with small testing programs and a neat media analyser.
MiniVideo has been tested with several CPU architectures (x86, SH4, MIPS, ARM).
The project uses a dual CMake/QMake build system (CMake is prefered though). Both library and test programs can be installed into your system.
MiniVideo has been initially developed in 2010/2011 during an internship I did in a French company called httv, as a small video decoding library developed from scratch in C. Its goal was to generate video thumbnails, with a source code easy to read and to understand for learning purpose. After a clean-up pass, the code has been published early 2014 with httv permission under the LGPL v3 license (video framework) and GPLv3 (test softwares).
The minivideo library can:
- Open video files with various container to demux and remux audios/videos content.
- Open H.264 compressed videos and decode/export intra-coded pictures.
- Extract various metadata from container and elementary streams.
- Map exact container structure to xml file / GUI.
Supported video codec (decoding)
- H.264 / MPEG-4 part 10 "Advance Video Coding"
- I frames only...
- please note that at still a few bugs inside CABAC decoding process being worked on...
Supported container formats (import modules)
- AVI [.avi]
- WAVE [.wav]
- ASF [.asf, .wma, .wmv]
- MKV [.mkv, .webm, ...]
- MP4 / MOV (ISOM container) [.mp4, .mov, .3gp, ...]
- MPEG-PS (MPEG "Program Stream") [.mpg, .mpeg, .vob, ...]
- MPEG-1/2 "elementary stream" [.mpg, .mpeg]
- H.264 / H.265 "elementary stream" ("Annex B" format) [.264, .265]
- MP3 "elementary stream" [.mp3]
Supported container formats (export modules)
- Elementary Streams
Supported picture formats (output modules)
- jpeg (internal OR when libjpeg support is available)
- png (internal OR when libpng support is available)
- webp (when libwebp support is available)
- bmp
- tiff
- tga
Building minivideo library
$ cd minivideo/build/
$ cmake ..
$ make -j$(grep -c ^processor /proc/cpuinfo)
You can change several build options directly into the "minivideo/CMakeLists.txt" file, or you can tune CMake by adding extra arguments to the cmake call:
-DCMAKE_BUILD_TYPE=Release/Debug
-DCMAKE_BUILD_Mode=Dynamic/Static
-DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/xxx.cmake
-DCMAKE_INSTALL_PREFIX=/usr/bin
Installation into the system, available for root user:
$ su
$ make install # system installation, need root user
Generating online documentation with Doxygen
$ cd minivideo/doc/
$ ./generate_doxygen.sh
Open "minivideo/doc/doxygen.html" with your favorite browser.
Generating error report with cppcheck
$ cd minivideo/doc/
$ ./generate_cppcheck.sh
Open "minivideo/doc/cppcheck.html" with your favorite browser.
Building MiniVideo's testing softwares
Do not forget "FindLibMiniVideo.cmake" directory in the cmake/modules/, which defines how to find the library (libminivideoframework.so file) and its header (minivideoframework.h file).
In case of problem, it may be necessary to manually indicate the paths of these files.
$ cd mini_analyser/
$ qmake
$ make
$ cd mini_extractor/build/
$ cmake ..
$ make
$ cd mini_thumbnailer/build/
$ cmake ..
$ make
Installation into the system, available for root user with both testing softwares:
$ su
$ make install # system installation, need root user
Using mini_analyser
$ cd mini_analyser/build/
$ ./mini_analyser
Then drag and drop files to analyse them!
Using mini_extractor
$ cd mini_extractor/build/
$ ./mini_extractor -i 'myfilepath' [-o 'mydirectory'] [-a nb_tracks] [-v nb_tracks]
Command line arguments:
-h : print help
-i : path to the input video
-o : path to the output folder, where extracted streams will be saved
-a : maximum number of audio stream(s) to extract
-v : maximum number of video stream(s) to extract
Using mini_thumbnailer
$ cd mini_thumbnailer/build/
$ ./mini_thumbnailer -i 'myfilepath' [-o 'mydirectory'] [-f picture_format] [-q picture_quality] [-n picture_number] [-m picture_extractionmode]
Command line arguments:
-h : print help
-i : path to the input video
-o : path to the output folder, where generated thumbnails will be saved
-f : export format for the thumbnails (can be 'webp' 'jpg' 'png' 'bmp' 'tga' 'yuv420' 'yuv444')
-q : thumbnail quality (1 to 100 range)
-n : number of thumbnail to generate (1 to 999 range)
-m : extraction mode for the thumbnails (can be 'unfiltered', 'ordered' or 'distributed')
