Intel introduced oneVPL support for the video processing and acceleration library the company includes in the FFmpeg oneAPI toolkit. The oneVPL video processing library facilitates invoking the latest 12th Gen Core Alder Lake and Intel Xe hardware in processors and in-house GPU acceleration. The acceleration is focused on Intel Arc Graphics and DG2 hardware, guiding the media SDK to support previous generations of graphics cards, with the ability to accommodate other potential CPU and GPU backends.
The new oneVPL support in FFmpeg will work in tandem with the VA-API or Video Acceleration API support in FFmpeg and in Intel GPUs.
As part of Intel’s long-standing contribution to the open source industry standards FFmpeg and GStreamer for media processing, Intel has added support for the oneAPI Video Processing Library (oneVPL) to FFmpeg. This is an important change because oneVPL is Intel’s evolution of the Intel Media SDK (Media SDK). Going forward, the primary way to access new GPU media features is through oneVPL (officially no longer adding features to the Media SDK), so the company encourages all Intel GPU media users to switch to oneVPL to take advantage of hardware features as soon as they become available.
Intel has put FFmpeg’s oneVPL integration in a “Cartwheel” repository, where the company collects and organizes a queue of any patches or bug fixes before being fully uploaded upstream.
Some patches cover VA-API, QSV and more parts. The company still has a number of updates in the repository that are currently in development and awaiting full implementation in the oneVPL integration. These updates will affect Intel Arc Alchemist and DG2 GPUs, but there is currently no specific date for any integrations to be uploaded to the main queue.
Readers can check out the latest developer article on Intel’s official webpage explaining how the company uses oneVPL in FFmpeg to get satisfactory streaming performance on the company’s GPUs.