Google releases libaom 3.7 AV1 encoder with more optimizations

[mayconhvab]

Following this week's releases of Intel's SVT-AV1 1.7 encoder and the release of the libavif 1.0 AV1 image file format, Google engineers have released libaom 3.7 as the newest feature release for the AV1 encoding library. The libaom 3.7 release focuses on new codec interfaces, compression efficiency and perception improvements, as well as speedups and memory optimizations. There is also a new speed (–cpu-used 11) and other improvements. Therefore, Google releases libaom 3.7 AV1 encoder with more optimizations.

Some of the key takeaways from libaom 3.7 include:
Compression efficiency improvements
Improved VoD encoding performance
0.1-0.6% BDrate gains for encoding speeds 2 to 6
Improved rate control accuracy in VBR mode
RTC encoding improvements
Screen Content Mode: 10-19% BDrate gains for speeds 6 to 10
Temporal layered video mode, for speed 10:
2 temporal layers at low resolutions: 13-15% BDrate gain
3 temporal layers in VGA/HD: 3-4% BDrate gain
Acceleration and memory optimizations
Speed up VoD encoder
2-5% for encoding speed 2 to 4
9-15% for encoding speed 5 to 6
ARM
Standard bit depth
speed 5: +31%
speed 4: +2%
speed 3: +9%
speed 2: +157%
High bit depth
speed 5: +85%
RTC Acceleration
Screen content mode
15% IC acceleration for speeds 6-8
ARM: 7% for speed 9, 3% for speed 10
Temporal layered video mode
7% speedup for 3 temporal layers in VGA/HD, to speed 10
Single Layer Video
x86: 2% IC speedup for speeds 7-10
ARM: 2-4% acceleration at speeds 5-10
It's great to see that all the AV1 encoding optimizations don't take anything away.