VVenC & VVdeC H.266 open source video encoder and decoder work on x86 and Arm

VVenC and VVdeC are open-source software H.266/VCC video encoder and decoder respectively that are optimized to use SIMD instructions on x86 (SSE42/SIMDe and AVX2) and Arm, and the decoder runs on Windows, Linux, macOS, and Android.

H.266, aka VCC (Versatile Video Coding) video compression standard was adopted in 2020 promising to reduce data requirements by around 50% compared to the previous H.265/HEVC standard at the same visual quality. H.266 should also outperform the royalty-free AV1 video codec. We hadn’t seen news since the announcement, but this may be changing with the Realtek RTD1319D processor unveiled with support for both 4K H.266 and AV1 video decoding last September, and progress made on the VVenC & VVdeC H.266 open-source software encoder/decoder as been discussed during FOSDEM 2023.

VVdeC VVenC H.266/VCC decoder & encoder

The Fraunhofer HHI group has been working on VVdeC and VVenC since the specifications were finalized in 2022. Both are based on VTM reference software for VCC, are written in C++ with a pure C interface, implement vectorization without relying on an assembler, and are offered under a BSD 3-Clause Clear license that explicitly does not grant any patent rights with the code available on GitHub.

VVdeC is fully compliant with the Main10 profile, can utilize over 30 threads, and works on Windows, Linux (x86, Arm, RISC-V…), macOS (x86 and Arm), as well as Android. Memory useage has been reduced by three times since the first release, and the developers continue making small improvements.

H.266 video decoder windowslinux macOS android

The VCenC open-source H.266 encoder is optimized for VOD (Video-on-Demand) and offline operation and comes with five presets: faster, fast, medium, slow, and slower, each providing a different balance in terms of quality and encoding speed. The encode supports multi-threading but does not work with more than 32 threads at this time. The developers further expect to work on software efficiency. Encoding works faster on the Apple M1 Arm chip compared to an Intel Core i9-12900H processor when using 8 threads.

H.266 Arm vs Intel

VVenC & VVdeC can already be integrated into FFmpeg using 3rd party patches, which then allow integration into mpv, VLC, and ExoPlayer. Finally, there’s also the VVdeC Web Player using WebAssembly to play H.266 videos (without audio) in your browser.

 

Further information may be found in the presentation slides and the FOSDEM 2023 video embedded above where H.266 vs AV1 is shortly discussed at the end of the video.

Via Phoronix

Share this:

Support CNX Software! Donate via cryptocurrencies or become a Patron on Patreon

ROCK Pi 4C Plus
Subscribe
Notify of
guest
The comment form collects your name, email and content to allow us keep track of the comments placed on the website. Please read and accept our website Terms and Privacy Policy to post a comment.
3 Comments
oldest
newest
Anonymous
Anonymous
1 year ago

The slapfight on Phoronix about this is worth checking out. 105 comments and counting.

I hope we see AV2 soon as an answer to this.

Bill
Bill
1 year ago

If the adoption of AV1 is anything to go by, I suspect gat this one may become adopted in abut 5 years time when we here about the new H268 variety that will require yet more new hardware.

Tim
Tim
1 year ago

This mirrors what I was going to say so get out of my head 😂

Honestly AV1 is impressive however and I’m glad my phone has hardware decode even if I had to find and download samples just to test the functionality in 2023…. :/

Khadas VIM4 SBC