We work on Linux opensource video and AI software for Intel

Intel Media and Audio for Linux contributes/maintains a family of media software, including libVA, gstreamer-vaapi, gstreamer-msdk, , FFmpeg-vaapi, FFmpeg-qsv, ffmpeg-DNN, libxcam and SOF. libVA is an open source software library and API specification to provide access to hardware accelerated video decoding/encoding and video processing. Gstreamer-vaapi is a GStreamer plugin that allows hardware accelerated video processing through libVA. Gstreamer-msdk is a Gstreamer plugin for the Intel® Media SDK to execute video hardware acceleration.FFmpeg-vaapi supplies hardware acceleration based on the low-level VAAPI interface that takes advantage of the industry standard VA API to execute high-performance video transcoding. FFmpeg-qsv supplies hardware acceleration based based on Intel Media SDK library. FFmpeg-DNN (deep neural network) module is an FFmpeg component that supports dnn-based filters. libxcam is an open source camera library to support computational photography features like 3A and image processing. SOF(Sound Open Firmware) is an open source audio DSP firmware and SDK that provides audio firmware infrastructure and development tools for developers who are interested in audio or signal processing on modern DSPs. All of these projects increase the speed and performance of decoding and encoding compressed digital video, imaging and audio. They also support workload offloading of CPU.

These projects are provided for operating system vendors, systems integrators, and original device manufacturers who are creating branded or customized operating systems for embedded devices, phones, tablets, convertibles, desktops, gaming and entertainment systems and more. Independent software vendors that create applications and games can also take advantage of the freely available source code and binaries.

lightbulb_outline View ideas list


  • ffmpeg
  • gstreamer
  • libxcam
  • vaapi


mail_outline Contact email

Intel Video and Audio for Linux 2021 Projects

  • Shubhanshu Saxena
    Async Support for TensorFlow Backend in FFmpeg
    This project focuses on implementing an asynchronous mechanism for model inference and batch execution in the TensorFlow backend of the FFmpeg Deep...
  • Eugene Lyapustin
    Equirectangular Projection (ERP) and Cubemap Projection (CMP) conversion
    360° video is getting popular during the past years and it is predicted to evolve [1]. Compared to virtual and augmented reality, 360° video is much...
  • Hongcheng Zhong
    VVC Parser for Gstreamer
    Versatile Video Coding (VVC) is the latest video compression standard after HEVC. GStreamer need to support VVC, and we should start from a VVC...