And I'd rather watch my release than one with this encoder and settings

It's well known that aom, while slow, is the best AV1 encoder because it provides better quality at the same size. If you have a source that says otherwise, feel free to post a link. Also, what specifically is bad about those parameters? You are free to point out what I should change so I can test it. I already tested many configurations a year ago and found that these gives the best size for a given quality. Maybe something has changed in the meantime, but I don't think so. Again, just specify which parameters or values you dislike so I can test them, instead of merely saying you don't like them without giving reasons. I too am able to go to your encode and say they suck without giving reasons.

It's well known that aom threads poorly and that av1an is meant to mitigate that by running encoder processes in parallel. Then why are you disabling chunking and leaving the encoder set keyframes at constant intervals? This is extremely confusing. What kind of person claim their encoder and settings are the most efficient without being part of the communities that regularly test and confront tools to always stay updated on what's best at a given time? You simply appear as ignorant by claiming something that is apparently well known according to you, but is actually outdated information that you've likely read on reddit or some other incompetent forum. My source is the community of encoders that dominates the AV1 releases sphere and tries to push innovation forward. What you should change isn't just a setting or two, but the entire way you envision video encoding, assuming you want to do serious stuff here.

So, for starters, spawning a single aomenc instance with av1an is borderline useless. You should not let the encoder place keyframes at all if you're gonna make use of av1an's scene detection. However with a min length that high, you are bound to be missing lots of scene changes, meaning lots of keyframes won't be placed at the right place. Some of your settings don't make sense following your "these gives the best size for a given quality" statement, because I have myself proven otherwise using my normalized methodology, the one I have showcased for instance in the [SVT-AV1 v1.8.0](https://wiki.x266.mov/blog/svt-av1-deep-dive) blog post I wrote. Furthermore, you are using a broken encoder and you probably didn't even notice it. `--enable-keyframe-filtering=1` in [mainline aomenc](https://aomedia.googlesource.com/aom/) has been broken for years, leading to blocking and other kind of annoying distortion on keyframes, which affects the neighboring frames as well as a consequence. Not only that, but forks of aomenc have existed for years as well, which on top of fixing some dubious behaviors like the keyframe-filtering=1 bug offer some psychovisual tuning in order to visually improve quality at a given filesize in ways that the original encoder was never designed to. I can mention [aom-av1-psy](https://github.com/BlueSwordM/aom-av1-psy/) which is now outdated but served as a base for [aom-av1-lavish](https://github.com/Clybius/aom-av1-lavish) and [aom-av1-psy101](https://gitlab.com/damian101/aom-psy101), which are the go-to aom-based encoders to use nowadays if you care about quality even slightly.

All of that doesn't mention that [SVT-AV1](https://gitlab.com/AOMediaCodec/SVT-AV1/) has now surpassed all aom-based encoders simultanously in efficiency and speed for a while, and that a psychovisual fork of SVT-AV1 has also appeared since, sporting saner defaults, new features and better documentation. It's called [SVT-AV1-PSY](https://github.com/gianni-rosato/svt-av1-psy/) and over at [AV1 weeb edition](https://discord.gg/83dRFDFDp7), we're helping testing and maintaining it. Keeping yourself informed should be the strict minimum to not appear clueless. We're all far from perfect, but at the very least we can patch each others weaknesses and flaws if we work together towards the same goal of spreading AV1 greatness over the anime community.

hey at least he uploads at 1080p no way anyone is going to sit back on a 4k tv and watch 720p thanks for the upload

At least one could make the effort to make a release worth downloading.

thanks

[New 1080p AV1 release @2.9 GiB](https://nyaa.iss.one/view/1854203) > Compared to the [Quantum] release, mine got: > - higher bitrate audio > - better detail retention most of the time, due to using a better source as a starting point and drastically better encoding settings > - higher appeal because not riddled with banding, properly antialiased,... [Banding situation pictured](https://slow.pics/c/13SkdQLv) > - proper video color metadata set, to avoid bad color rendition with common media players > - more subtitles, a cover > - similar decoding speeds while using the computationally intensive grain synthesis feature > And despite of that, the release is overall slightly smaller.