SVT-AV1 VS encode-scripts

The docs mention that the unit for this option is in "frames" (not "seconds").

From the pull request:

In these tests SVT-AV1 beats x265 on quality:

- FullHD: http://compression.ru/video/codec_comparison/2022/main_repor...

- FullHD 10-bit: http://compression.ru/video/codec_comparison/2022/10_bit_rep...

- 4K: http://compression.ru/video/codec_comparison/2022/4k_report....

SVT-AV1 has seen a number of speed ups in recent releases:

https://gitlab.com/AOMediaCodec/SVT-AV1/-/releases

I don't see this options under the "Advanced features" like for H.265/HEVC (for example). Also, adding "-input-depth 10" manually to the command line doesn't work (it doesn't recognize the parameter, despite it clearly being written here as one of them).

Docs for SVT-AV1 at https://gitlab.com/AOMediaCodec/SVT-AV1/-/tree/master/Docs

I find it interesting that you know the project well enough to get to the pre-compiled binaries page, yet somehow missed that the merge request explictly calls out a VMAF LOSS for the SSIM tune by several percentage points.

SVT-AV1 has no NEON.

I used the precompiled FFmpeg from SVT-AV1's git: https://gitlab.com/AOMediaCodec/SVT-AV1/-/jobs/4540276837

This'll be long-winded excitement to talk about the weird little community, but I think a lot of that depends on the circles you run in and the content you consume.

There is surely a lot of low-effort GUI handbrake encodes online. But most of the """well-respected""" piracy groups put a surprising amount of effort into filtering and such to correct artifacts, both due to the compression and due to the source material itself.

A lot of these people are using tools like VapourSynth with a variety of scripts they've put together and x264 or x265 directly rather than ffmpeg. You can see a couple of guides written about some of the processes they perform:

- <https://silentaperture.gitlab.io/mdbook-guide/introduction.h...> (all video content)

- <https://guide.encode.moe> (anime-focused)

And some links to the kinds of filtering code pirates write for movies/tv/anime:

- <https://git.concertos.live/OpusGang/EncodeScripts> (MANY VapourSynth and AviSynth scripts for both live-action content and anime)

- <https://github.com/Beatrice-Raws/encode-scripts>

And while not directly related to the encoding side of things, but if any of that is interesting, in addition to the encoding side of things, pirate fansubs also get pretty complex, particularly for anime since, unlike the unstyled SRT subs most people come across for foreign movies online, anime fansubs tend to use ASS [1] subtitles with lots of styling to accomplish things like cleanly replacing Japanese text in a letter someone is reading or adding non-distracting subtitles for background text (e.g., signs on buildings, etc).

To do a lot of that, though, these subtitles often pack fonts into the video container to allow the media player to render things as expected without resorting to "hardsubbing" (i.e., pre-rendering the subtitles into the video itself)—which is one of many reasons container formats like Matroska (MKV) is so popular in those communities.

An interesting thing to see come out of that is that I have noticed some fansubbing groups move to proper build tools, like Gradle, to automate portions of their workflows. As an example, SubKt, a Gradle plugin, allows them to essentially have CI/CD for their subtitling projects by doing integrity checks on the fonts, linting the subtitles/fonts to ensure the selected fonts actually have glyphs for all the text, templating and merging so that different team members can work on things like the script/timing while another does styling, and then packaging and publishing tasks to bundle everything up into an MKV at the end and upload the result to torrent sites.

If any of that is interesting, here are some links to SubKt + some real-world finished projects making use of it:

- <https://github.com/Myaamori/SubKt>