/tech/ - Technology★

>>930263 (OP)

1. AV1 looks better than h265/HEVC

2. AV1 has more of a sandbox decoder allowing encoding quality to still rise after bitstream freeze (for newly encoded things of course)

3. Hardware support isn't that important with modern powerful processors and will come with the content.

Also wish to remind you that there's an impenetrable wall between nonapplefags and applefags preventing acceptance of things like HEIF, the newest USB alternative, Retina Displays, HiDPI stuff, gay mouses and keyboards with switched ctrl, super, alt keys.

4. h264/AVC is an old pile of shit (the best of it's time tbh) and is just there because otherwise most videos on the internet would need to be reencoded losing quality.

5. It seems to have a lossless encoder

6. There's also an image format based on it (AVIF) which could spread contrary to BPG (based on h265).

7. Most important source of videos today is THE WEBBROWSER and guess how much of this sector runs chromium. Firefox gets it too and the rest simply has to comply.

YES, those stupid cat videos will help AV1 becoming popular!

>AV1 reference encoder runs at the speed of olympic molluscs with a 10% chance of shitting out a valid bitstream?

Yes. It still seems to be unfinished and unfrozen.

Freeze soon™

>what would you put in

AV1 when it's finally finished tbqh

>>930428

>Most important source of videos today is THE WEBBROWSER

This brings up the question: Why bother making something decode fast if most of the load will come from the browser anyway?

It's like how all over the world governments start metropolitan "save water" campaigns during droughts even though big farms generally consume 97% of the water supply of countries self sufficient in farm produce, and shamelessly waste it.

Also contrary to h265 it will be allowed in the WEBM subset of Matroska baking it into the web. Basically it has already won even through it's still unfinished. You may call it an unfair battle but at least the royalty fags didn't win.

>>930429

>Why bother making something decode fast

AVIF images, "GIFs", TV, phones, ...

I'm yet to see a high quality video encoded with HEVC, videos encoded with it always seen to have visible loses in quality.

>>930449

>easter island stonemasonry BBSes

Intrigued. Where can I sign up?

>>930470

You can find it this way >>>/pol/

>>930428

You are the most worthless poster I think I have ever seen. Bitstream is frozen and it still takes a year (not hyperbole) to encode one fucking movie. By the time they have hardware encoders you will have one fucking movie done. It's a secretive house fumigation where nobody was told and everyone is dead inside. There are no survivors. h.266 will be done when av1 is "ready". Absolutely shit.

>>930478

FUCKING DOGSHIT

>>930263 (OP)

>muh hardware CODEC

Isn't Vulkan/D3D12 supposed to mostly obviate that?

>If you could design your own ideal digital video coding standard regardless of patent/copyright infringement what would you put in it /tech/?

Fractal coding, throw this outmoded macroblock shit in yhe trash. Also, make sure hierarchical bitrate peeling is supported, plus optional lightfield coding is there too.

>>930449

>But the codec spec is finished

It's not, it was a PR stunt (and you fell for it).

>>930428

>YES, those stupid cat videos will help AV1 becoming popular!

Braindead people first have to stop using GIF for that.

GIF is 30 times less efficient than either VP9 or H.264, yet these fuckers still use it, go figure.

>>930428

Why do you think Google/YouTube has the ability to transcode every fucking video to VP9, yet it takes literally months to encode any "HD" video with libvpx?

If the format is open, but all the good encoders are not available to the general public, and thus only one company is actually able to reap the benefits of the format, is it worth it? How is it better than H.264, which has x264 which is efficient and free?

>>930658

>Braindead people first have to stop using GIF for that.

That's already happening. For example on Twitter.

NOT BECAUSE THE RETARDS REENCODE IT BUT BECAUSE THE SITES DO IT FOR THE SERVER PERFORMANCE AND FILESIZE

>>930662

>How is it better than H.264

Encode a video in H264 and one in VP9 and tell me which of them looks better at same filesize. Filesize is the reason why we don't just fill a folder with PNGs and FLAC files, all slapped together by XML textfiles standardized by W3-jews!

>takes literally months

>literally months

>literally

>HD aka 720p

>months

WTF are you talking about? Takes some hours for a movie but not months.

>>930654

Would agree if AV1 was finished.

>>930478

https://aomediacodec.github.io/av1-spec/

>Last modified: 2018-06-14 12:54 PT

>Draft Document

>This is a draft document that will change substantially before release. Do not rely on its current contents. To comment on the document, use the issue tracker.

Learn reading instead of repeating announcements we already know about.

see >>930557 (you did a good job pointing it out)

VP9 changed substantially even after it's "release".

For example the addition of an alpha channel. That's why Chrome was so ahead because Google basically made it while implementing it. It's not better than HVEC but royalty free and better than H264.

>>930662

http://wiki.webmproject.org/ffmpeg/vp9-encoding-guide

Here's a guide. Maybe you missed some stuff like how 2 pass works or -threads 8 to make use of your multi core cpu.

>>930753

I can play the video in palemoon just fine. Firecucks btfo.

>>930795

>>930808

Can mpv reliably playback AV1 webms these days?

Would be a nice workaround for the lack of native support in bowsers at the moment it'd also be nice if 8ch would recognize V_AV1 as a valid codec for webbums.

>>930429

>Why bother making something decode fast if most of the load will come from the browser anyway?

Not all videos are 360p, you know? Try to compare that to playing 1080p60 video (w/o GPU decoding) before spewing retarded bullshit.

>>930691

>NOT BECAUSE THE RETARDS REENCODE IT BUT BECAUSE THE SITES DO IT FOR THE SERVER PERFORMANCE AND FILESIZE

An the quality is total shit, because 2 generations of different kinds of losses.

They got what they deserved, but still didn't understand a thing.

>>930691

>Encode a video in H264 and one in VP9 and tell me which of them looks better at same filesize

AND WITH THE SAME AMOUNT OF CPU TIME SPENT

If I try to make libvpx encode with similar speed to x264 -preset slower, the end result is crap, obviously worse than x264.

>>930691

>WTF are you talking about? Takes some hours for a movie but not months.

what kind of a supercomputer are you using?

>>930835

Install gentoo, I can encode a 3 hour video as webm with 1 core used in 24 hours.

>>930826

You can just tell most browsers to open mp4/webm/whatever files with a program of your choice. I chose mpv.

>>930879

That's quite retarded, since you must download it in full before viewing it. It's NOT the same as passing the URL to mpv.

>>930478

>h.266 will be done when av1 is "ready".

Do you think that h.265 will have support outside of specialized programs by then?

>>930658

The reason for that is GIF is seen as an image format and not a video format.

Because it's seen as a image format it becomes more portable for normalfags to use.

The only real thing that could replace GIF is Animated pngs, but retards and lazy devs still haven't implemented them properly.

>>930700

Just ignore faggots who are desperate to fit in.

>>931305

Also apng isn't a standard and you can't easily tell which png contains apng data.

I hope for AVIF.

>>931305

>The reason for that is GIF is seen as an image format and not a video format.

>Because it's seen as a image format it becomes more portable for normalfags to use.

In both cases it's a fucking file and it's equally portable.

>>931305

>Animated pngs

>>931305

>retards and lazy devs still haven't implemented them properly

because it's still shit for video compression (although supporting more than 256 colors)

if one wants to encode a piece of near-photorealistic footage, deliberately throwing 25+ years of progress in video encoding techniques development out of the window is 100% insane.

>>930826

>windows

>>>/bog/

>>930263 (OP)

Could a full VP9 decoder or encoder (like in the Kaby Lake iGPUs) work as a hybrid AV1 decoder/encoder with a software update (like the VP9 decoder/encoder in the Skylake iGPUs)?

>>931561

>>930753

>>931277

cpu-used 0 in VP9 is for extreme qualityfags only. 1 is a lot faster with only slightly lower quality, while everything above 1 is inefficient (should just go ahead and use x264 instead if you were going to use a cpu-used value above 1)

>>931661

A partial HW decoder via software update might be possible as AV1 is sort of an extended fork of VP9, but any sort of quarter-decent AV1 HW encoding will definitely require a custom ASIC.

Judging by how unreliable the current reference decoder/encoder are we're probably at least 2-3 years from these sorts of thingsgiving MPEG and friends more time to sort out their patented HEVC licensing faggotry.**

>>931108

Try opening a h.265 video with a browser (like most non-/tech/fags would).

>>931108

What kind of outdated piece of shit are you running that can't even render the up/down symbols?

>>931867

he's blocking fonts