NVIDIA has added support to the open-source FFmpeg multimedia library to allow for the advantages of accelerated AV1 video encoding for compatible GPUs through the VDPAU API while utilizing the current-generational NVIDIA RTX 30 graphics cards that use the Ampere architecture.
NVIDIA receives AV1 VDPAU hardware acceleration with the updated FFmpeg multimedia library for use in compatible GPUs
The VDPAU API, also known as the Video Decode and Presentation API for UNIX, provides an interface to video decode acceleration and presentation hardware currently in contemporary graphics cards.
With the new support received by NVIDIA for AV1 VDPAU decoding for accelerated hardware, the company has the accessibility of the FFmpeg software libraries. FFmpeg is a compilation of software libraries that can document, transform and stream digital A/V in multiple configurations. FFmpeg possesses the libavcodec audio/video codec library and a decoder/encoder codec-suit that comprises the foundation of numerous free and open-source software multimedia applications. The AV1 video decoding functions concur with libvdpau 1.5+ and employ an Ampere-based or higher graphics card with the essential hardware abilities.
NVIDIA GPUs possess a hardware-based decoder (referred to as NVDEC) which delivers ultimately accelerated hardware-based video decoding for multiple, widespread codecs. After introducing FFmpeg 4.4, the AV1 decoding capabilities have been present on NVIDIA GPUs through the NVDEC “NVIDIA Decode” interface included in the company’s Video Codec SDK. FFmpeg also helps with Intel’s Quick Sync Accelerated Video Decoder (also known as QSV-accelerated decode), the Windows DXVA2/D3D11VA AV1 decode, and processor-based decoding through the DAV1D project.
The Windows DXVA2, DirectX Video Acceleration is an API and an interconnected DDI for operating hardware acceleration to boost video processing. Software video processors and codecs can employ DXVA to offload certain CPU-heavy operations to the graphics processor.
With the current FFmpeg Git, merged last Saturday and located here, AV1 VDPAU decode is prepared to see use in determining VDPAU to the more recent NVIDIA NVDEC hardware-based decoder. When not seen in NVIDIA products, the AV1 VDPAU openness may enable the Mesa Project’s Gallium3D drivers that currently support the VDPAU video acceleration state tracker as more graphics processors update to the new AV1 decode hardware capability.
News Sources: Phoronix, Github,