After introducing dav1d at the Video Developer Days 2018, the team behind dav1d announced yesterday that they will be releasing its first version very soon. This version will come with features like Film Grain, Super-Res, Scaled References, and other more obscure features of the bitstream.
dav1d is a new AV1 cross-platform open source decoder, which is developed with speed and correctness in mind. Currently, the reference decoder (libaom) of AV1 needs major improvements. libaom was basically developed for research purposes, therefore the VideoLAN, VLC, and FFmpeg communities started working on this new decoder, sponsored by the Alliance of Open Media (AOM).
dav1d will come with features such as Film Grain, Super-Res, Scaled References, and other more obscure features of the bitstream, which will be available for both 8 and 10 bits depth. Along with that, the developers have improved the public API and the reduced the size of C code.
To assure secure decoding for AV1, the team has used fuzz testing. “Then, we’ve fuzzed the decoder a lot: we are now above 99% of functions covered, and 97% of lines covered on OSS-FUZZ; and we usually fix all the issues in a couple of days. This should assure you a secure decoding for AV1,” Jean-Baptiste Kempf, the president of the VideoLAN non-profit org said in his announcement.
dav1d shows amazing performance on AVX2 processors, which means that it covers more than 50% of the CPUs used on the desktop. The developers are working on the SSE and ARM optimizations, which will be completed in the next few weeks.
To test its performance, clips were taken from Netflix, Elecard and YouTube and was compared with aomdec. When tested on Haswell it gave an average 2.49x relative decode performance. When tested on a more modern Zen machine, dav1d gave even higher average relative decode performance at 3.49x.
Here’s how the global comparison looked like:
You can read more in detail about dav1d on Jean-Baptiste Kempf’s announcement.
Presenting dav1d, new lightweight AV1 decoder, by VideoLAN and FFmpeg
YouTube starts testing AV1 video codec format, launches AV1 Beta Playlist
A new Video-to-Video Synthesis model uses Artificial Intelligence to create photorealistic videos
I remember deciding to pursue my first IT certification, the CompTIA A+. I had signed…
Key takeaways The transformer architecture has proved to be revolutionary in outperforming the classical RNN…
Once we learn how to deploy an Ubuntu server, how to manage users, and how…
Key-takeaways: Clean code isn’t just a nice thing to have or a luxury in software projects; it's a necessity. If we…
While developing a web application, or setting dynamic pages and meta tags we need to deal with…
Software architecture is one of the most discussed topics in the software industry today, and…