https://gitlab.com/AOMediaCodec/avm/-/commit/6c74d9b1b2566e9b0c735dcb1a21858f24a00298

C076_inter_mode_ctx C076 inter mode contexts

https://gitlab.com/AOMediaCodec/avm/-/merge_requests/781 https://gitlab.com/AOMediaCodec/avm/-/issues/166

Tutorial to how encode a video https://www.nxrte.com/jishu/11119.html#group=nogroup&photo=0

01/06/2023 18:40 https://www.oxfordlearnersdictionaries.com/definition/english/trade-off_2?q=trade-off

01/06/2023 https://www.bloomberg.com/press-releases/2023-06-01/alliance-for-open-media-s-volumetric-visual-media-working-group-opens-call-for-proposals-on-static-polygonal-mesh-coding The important part is this High quality of the 3d graphics Will be downloaded here Will be they anonymezed until Junpstart Other words are just fillers

https://cloud.tencent.com/developer/user/1601916 tencent news about codecs

An Open Video Dataset for Screen Content Coding https://ieeexplore.ieee.org/document/10017995 https://cloud.tencent.com/developer/article/2260936?areaSource=&traceId= Y. Wang Daede 2022 IEE

https://codecwar.com/compare/summary?id=07f9479cf20536da0e46190033e71e08 Comparison page Codecs Comparative Report

https://www.nxrte.com/jishu/6022.html#group=nogroup&photo=0 8K OTT Source: Coal Mine Factory Technical Articles

https://ca.linkedin.com/in/danilovucetic

Software Engineering Intern, Masters Google

Sep 2022 - Dec 20224 months

Mountain View, California, United States

Researched lightweight deep learning approaches for video restoration in the AV2 codec, generating multiple models capable of increasing quality (PSNR gain) by up to 2.5%.

Developed a method to generate quantized weights for linear combinations of output channels, called encoder-guided restoration, leading to improvements in codec performance of up to 2%.

Investigated distillation as a method of achieving highly efficient and effective models for video restoration.

Explanation to VVC https://www.nxrte.com/jishu/10757.html#group=nogroup&photo=0 Codec • Source: Technical Articles

Evolution of codec, Ali266 unlocks new growth in the field of ultra-high-definition video Codec • Source: Video Cloud Technology • February 1, 2023 at 5:04 pm • Technical Articles

https://bloggeek.me/rtcscale-2023/#h-niklas-enbom-meta-av1-for-meta-rtc Sriram Srinivasan / Meta – Real-time audio at Meta Scale audio WebRTC Niklas Enbom / Meta – AV1 for Meta RTC An integrator's perspective. Meta will adopt AV2

livevideostack in the early 2017 https://www-livevideostack-cn.translate.goog/news/2023-codec/?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=it&_x_tr_pto=wapp https://vcgit.hhi.fraunhofer.de/ecm/ECM https://sh2022.livevideostack.cn/

I tried to run Av1an docker image on my Ubuntu 22.04 machine with the following:

docker run --privileged -v "$(pwd):/videos" --user $(id -u):$(id -g) -it --rm masterofzen/av1an:latest -i input.mkv

However, the following error occured:

thread 'main' panicked at 'Input file "input.mkv" does not exist!', av1an-core/src/lib.rs:1206:5

The current directory in the terminal is the directory where input.mkv is stored in. So running pwd should have mounted the directory with the input.mkv to /videos. I also set mode for input.mkv to 777 just in case.

Credit asm-c I didn't even know the AOM had a working group for Volumetric Visual Media. Asm-c At first I thought this might be related to the light field stuff that's in libaom (think sci-fi holograms), but it's actually about compressing "traditional" 3D models. Sounds good to me, though I have no idea what the situation is when it comes to 3D model compression. I assume it's a hodgepodge of no compression (or just making lower-detail copies for LoD purposes and such) and maybe some general-purpose/proprietary compression algos depending on which software/engine is being used and what it's being used for. The first search result for "3d mesh compression" is something from Google, so I guess it makes sense that the AOM is interested.

Very important talk to know VMAF and understand AV1 and AOMedia codecs. On YouTube.com

Found it on: MediaTek Pentonic 1000, SoC for 4K smart TV

TVs from brands like Sony or Panasonic will offer four HDMI 2.1 inputs from 2023, thanks to the new MediaTek SoC designed for high-end 4K smart TVs Pentonic 2000

MediaTek has announced the Pentonic 1000 SoC, developed for high-end 4K smart TVs with 120 or 144 Hz panels. MediaTek processors are used by TVs from brands such as Samsung, Sony, Panasonic, Philips and Xiaomi, as well as Amazon and Roku streamers. The Pentonic 1000 integrates a 2.0 GHz quad-core CPU with 64-bit LPDDR4 / 4X RAM at 4,266 Mbps, a dual-core GPU and a dedicated AI processor. The new processor supports four HDMI 2.1 inputs, so even smart TVs from Sony and other MediaTek partners will be able to count on four HDMI 2.1 ports from next year instead of the usual two. The manufacturer said that the smart TVs equipped with the Pentonic 1000 will arrive in the first quarter of 2023.

It don't miss the new H.266 / VVC codec, heir to HEVC, as well as AV1, another of the new generation codecs. AI-Super Resolution 2.0 with "detail reconstruction" takes care of scaling the video sources of lower resolution. It continues with 4K / 144 Hz VRR support, HDR10 + Adaptive and HDR10 + Gaming formats, Dolby Vision IQ with Precision Detail, Dolby Atmos decoding and MEMC (Motion Estimation Motion Compensation or motion smoothing) technologies.. The first processor in the series was the Pentonic 2000, designed specifically for 8K TVs, followed by the Pentonic 700 for mid-range 4K TVs.

The Mediatek Pentonic 1000 chipset is a worthy addition to the Pentonic series. I think the ultra-fast Wi-Fi 6 setup in the 4k premium television segment. The said chipset is going to be for 4K 120Hz HDR displays and 4K 144Hz VRR which are specially for gaming. The SoC by the brand extends Dolby Vision IQ support along with a Precision Detail. The televisions with Pentonic 1000 are said to be launching in the first quarter of 2023. Also, for video streaming the chip supports - HEVC, AV1, and AVS3, along with the new VVC also known as H.266. I am eagerly waiting to see how users going to receive these enhancements in their 4K TVs.

Yes, I can't wait for 144Hz to be standard on all TVs. I hope all Samsung TVs will be 144Hz in 2023 including QDOLED. I nearly bought Samsung QN90B TV now but I will wait until Black Friday 2023 then it will have WiFi6 and 144Hz for larger sizes plus hopefully lots of other improvements and I hope the Samsung prices drops to match TCL.

Pentonic 1000 will be amazing for gaming performance and multimedia performance. The processor has a strong quad core structure with a maximum speed of 2.0GHz. The processor has 4 A73 Cores that work very fast and efficiently.

Doesn't support 4x 2.1. Still only 2x.

wavif A new AVIF encoder by Cédric Louvrier, French developer who made Pingo optimizer. For more information: https://en.wikipedia.org/wiki/AVIF

XnConvert (and XnView) could easily map them to a basic set of options like JPEG, but Pierre generally doesn't have time to do more than the bare minimum to throw a new format in. It's highly unlikely he'll ever add wavif, since there's already one AVIF encoder. The Jasper/Lurawave situation with JPEG-2000 is the only time I can remember two plugins overlapping to the same degree, and that was a disaster that's long excised.

How was the AOMedia AV1 idea. Which tools were under consideration after 0.0.1? What decoding complexity was expected? Why PVQ wasn't adopted? Is expected from AV2? What are alternative to the PVQ for optimizing rate distortion and get all the 13,8% HEVC gain?

Watch it all and at 34:00 sample images

Ok so after a full test with my 4k 24fps 30sec clip of Avengers infinity war (HEVC) we are with :

SVT-AV1 1.4.1 no grain denoise off: 19:18 minutes / 0.82 fps / 9.2% of source

SVT-AV1 1.4.1 grain synthesis 20 denoise off: 20:34 / 0.76fps avg / 9.2% of source ( -6.05% speed )

SVT-AV1 1.5 no grain denoise off: 15:20 minutes / 1.18fps avg / 9.356% of source ( +20.75% speed )

SVT-AV1 1.5 grain synthesis 20 denoise off: 16:09 / 1.09fps avg / 9.381 of source ( +16.11% of speed )

SVT-AV1 1.6 no grain denoise off: 10:57 minutes / 1.43fps avg / 9.168% of source ( +44.89% speed )

SVT-AV1 1.6 grain synthesis 20 denoise off: 11:43 / 1.33fps /9.192% of source ( +40.41% speed )

which means 20% speed boost between svt-av1 1.4.1 and 1.5. And another 30% between 1.5 and the new 1.6.

GG devs, you are amazing!

( this is with preset 3 crf 18 at 4k and the settings for grain when used, all in handbrake. mind that this is with a Ryzen 5800X3D with a little bit of downvolting for perf )

This is test post from kbin.social https://github.com/ernestwisniewski/kbin