Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Video Encoding > New and alternative video codecs

Reply
 
Thread Tools Search this Thread Display Modes
Old 13th June 2018, 03:42   #1101  |  Link
IgorC
Registered User
 
Join Date: Apr 2004
Posts: 1,298
Quote:
Originally Posted by Motenai Yoda View Post
Yep but then you'll find this

as they don't even complaing enabling alt-ref frames for the 2 pass test which give a HUGE quality boost, not even playing with qcomp.
So what their comparisons mean?
Maybe. I didn't know about that.
IgorC is offline   Reply With Quote
Old 26th June 2018, 10:19   #1102  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload:

VPx v1.7.0-541-gdd3d08f0c
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 20th July 2018, 15:32   #1103  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload:

VPx v1.7.0-653-g03e1bd397
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 31st July 2018, 16:41   #1104  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload:

VPx v1.7.0-747-g3ff77503c
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 9th August 2018, 16:02   #1105  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload:

VPx v1.7.0-798-gaab2aff9a (MSYS2; MinGW32: GCC 7.3.0 / MinGW64: GCC 8.2.0)
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 14th September 2018, 14:51   #1106  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload:

VPx v1.7.0-1002-gb3a837cbf (MSYS2; MinGW32: GCC 7.3.0 / MinGW64: GCC 8.2.0)
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 14th September 2018, 15:56   #1107  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 826
Quote:
Originally Posted by LigH View Post
New upload:

VPx v1.7.0-1002-gb3a837cbf (MSYS2; MinGW32: GCC 7.3.0 / MinGW64: GCC 8.2.0)
Any meaningful improvements these days?
mzso is offline   Reply With Quote
Old 14th September 2018, 16:05   #1108  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
A lot cleanups, fixes, improvements ... but I don't know any especially "meaningful" changes in detail, did not study this list.
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 23rd September 2018, 18:32   #1109  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload:

VPx v1.7.0-1064-g3448987ab (MSYS2; MinGW32: GCC 7.3.0 / MinGW64: GCC 8.2.0)
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid

Last edited by LigH; 23rd September 2018 at 18:37.
LigH is offline   Reply With Quote
Old 7th October 2018, 23:02   #1110  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload (MSYS2; MinGW32: GCC 7.3.0 / MinGW64: GCC 8.2.0):

VPx v1.7.0-1137-g4a47ef814
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 2nd November 2018, 14:07   #1111  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload (MSYS2; MinGW32: GCC 7.3.0 / MinGW64: GCC 8.2.0):

VPx v1.7.0-1288-g811759d86
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 13th December 2018, 16:18   #1112  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,787
New upload (MSYS2; MinGW32: GCC 7.4.0 / MinGW64: GCC 8.2.1):

VPx v1.7.0-1507-gc62d9d568
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 14th December 2018, 01:15   #1113  |  Link
utack
Registered User
 
Join Date: Apr 2018
Posts: 35
Netflix dropped a new article.
H264/H265/VP9 comparison using HVMAF with reference encoders and production encoders.
libvpx is doing mediocre'ish, EVE does well
Relative Bitrate savings for low and high quality using three test sets:



https://medium.com/netflix-techblog/...e-d45d0183ca95

Last edited by utack; 14th December 2018 at 01:21.
utack is offline   Reply With Quote
Old 14th December 2018, 02:53   #1114  |  Link
benwaggoner
Moderator
 
Join Date: Jan 2006
Location: Portland, OR
Posts: 2,794
Quote:
Originally Posted by utack View Post
Netflix dropped a new article.
H264/H265/VP9 comparison using HVMAF with reference encoders and production encoders.
libvpx is doing mediocre'ish, EVE does well
Relative Bitrate savings for low and high quality using three test sets:



https://medium.com/netflix-techblog/...e-d45d0183ca95
Now VMAF gives different scores for mobile, 1080p, and UHD screen sizes. They donít seem to specify which they used here.
Or perhaps this was done with an older VMAF implementation?

Iím having a hard time figuring out what settings they are actually using. It sounds like itís fixed QP, with one psychovisual parameter added for each ďperceptualĒ tuned mode. So are x264 and x265 fixed QP encodes with psy-rd=1? Is aq-mode=0? And really QP instead of CRF? And when aiming for PSNR, why not --tune psnr which is EXACTLY for that scenario! Psy-rd=0 is NOT tuning for PNSR!

And odd libvpx gets to use 2 passes while everything else is 1. Although that wouldnít really matter that much if itís truly a fixed QP encode with no rate control. But if using --tune psnr for x26? multipass encoding should help mean and harmonic mean PSNR.

This seems a quite poor study for predicting the real-world subjective quality different encoders can produce, as it will substantially underestimate the achievable perceptual quality the x26x codecs can deliver in the real world. Iíd expect just adding --tune film to x264 would improve VMAF a bunch in the perceptual case.

It sort of conflates encoder psychovisual tuning and bitstream capabilities. I imagine a port of x264ís psyovisual stuff to a vp9 encoder would offer big improvements, like was seen when x264 algorithms got ported into x265.
__________________
Ben Waggoner
Principal Video Specialist, Amazon Prime Video

My Compression Book
benwaggoner is offline   Reply With Quote
Old 14th December 2018, 05:34   #1115  |  Link
utack
Registered User
 
Join Date: Apr 2018
Posts: 35
Quote:
Originally Posted by benwaggoner View Post
Now VMAF gives different scores for mobile, 1080p, and UHD screen sizes. They don’t seem to specify which they used here.
>HVMAF is computed after scaling the encodes to the display resolution (assumed to be 1080p)

Quote:
Originally Posted by benwaggoner View Post
I’m having a hard time figuring out what settings they are actually using. It sounds like it’s fixed QP, with one psychovisual parameter added for each “perceptual” tuned mode. So are x264 and x265 fixed QP encodes with psy-rd=1? Is aq-mode=0? And really QP instead of CRF?
They are using the approach described here.
https://medium.com/netflix-techblog/...k-e19f1e3a277f
As far as I understood it encodes it "shot" multiple times, adjusting target bitrate until it hits a desired VMAF

Quote:
Originally Posted by benwaggoner View Post
And odd libvpx gets to use 2 passes while everything else is 1
It is, probably because everything else has lookahead?


Quote:
Originally Posted by benwaggoner View Post
And when aiming for PSNR, why not --tune psnr which is EXACTLY for that scenario! Psy-rd=0 is NOT tuning for PNSR!
They did tune for PSNR, in part1 "Results with the traditional approach"



Quote:
Originally Posted by benwaggoner View Post
This seems a quite poor study for predicting the real-world subjective quality different encoders can produce, as it will substantially underestimate the achievable perceptual quality the x26x codecs can deliver in the real world.
I don't think you can get any of the encoders to do much better, the article from above also goes into detail about how much the internal ratecontrol/quantization choice of all encoders could be improved with their "VMAF feedback loop".
utack is offline   Reply With Quote
Old 14th December 2018, 08:14   #1116  |  Link
benwaggoner
Moderator
 
Join Date: Jan 2006
Location: Portland, OR
Posts: 2,794
Quote:
Originally Posted by utack View Post
>HVMAF is computed after scaling the encodes to the display resolution (assumed to be 1080p)
That was the prior VMAF model. The new one also has a mobile and a UHD scale and comparison.

Quote:
They are using the approach described here.
https://medium.com/netflix-techblog/...k-e19f1e3a277f
As far as I understood it encodes it "shot" multiple times, adjusting target bitrate until it hits a desired VMAF
But thatís just tuning in BD rate. It isnít actually tuning any of the other parameters.

Also, VMAF itself has plenty of flaws as a metric. It doesnít do a good job differentiating between high quality encodes, doesnít pick up on issues with gradients well, and has other flaws. It is the best objective metric we have, absolutely. But real MOS subjective testing is required to be talking about percentage differences in BD-rate.

Quote:
It is, probably because everything else has lookahead?
Unless they are using a lookahead equal to clip duration, it wouldnít be apples to apples.

Quote:
They did tune for PSNR, in part1 "Results with the traditional approach"
No. If they were actually tuning for PSNR, they would have used --tune PSNR!

Quote:
I don't think you can get any of the encoders to do much better, the article from above also goes into detail about how much the internal ratecontrol/quantization choice of all encoders could be improved with their "VMAF feedback loop".
With fixed QP and no adaptive quant? I can do a LOT better than that with some tuning, and enocode a lot faster at the same time. A reasonably tuned - -preset slower would look better and be a lot faster than the --preset placebo without real psychovisual tuning they did here. By the description, I guess they turned off psychovisual features that are on by default, like aq-mode > 0. And fixed QP is a silly rate control mechanism, since the optimal average QP per frame varies a fair amount by the content in the frame. Anime needs lower QP than a jungle scene, for example (even though it can deliver lower QP at a lot lower bitrate than the higher QP of the jungle scene takes).
__________________
Ben Waggoner
Principal Video Specialist, Amazon Prime Video

My Compression Book
benwaggoner is offline   Reply With Quote
Old 14th December 2018, 16:37   #1117  |  Link
Beelzebubu
Registered User
 
Join Date: Feb 2003
Location: New York, NY (USA)
Posts: 47
Quote:
Originally Posted by benwaggoner View Post
I’m having a hard time figuring out what settings they are actually using. It sounds like it’s fixed QP [..] And really QP instead of CRF?
The paper says they use CRF for x264/5. I would recommend reading the paper if you want to know all the details, it's pretty detailed.

Quote:
Originally Posted by benwaggoner View Post
And odd libvpx gets to use 2 passes while everything else is 1. Although that wouldn’t really matter that much if it’s truly a fixed QP encode with no rate control.
The reason people don't use 1-pass CRF in libvpx is because it's pretty broken. Try it. Several bitstream (!!) features get *completely disabled* (!!) when using 1-pass encoding using libvpx. So you get several % BDRATE quality improvements when switching from 1-pass to 2-pass CRF.
Beelzebubu is offline   Reply With Quote
Old 14th December 2018, 19:01   #1118  |  Link
mandarinka
Registered User
 
mandarinka's Avatar
 
Join Date: Jan 2007
Posts: 705
Quote:
Originally Posted by benwaggoner View Post
It sort of conflates encoder psychovisual tuning and bitstream capabilities. I imagine a port of x264ís psyovisual stuff to a vp9 encoder would offer big improvements, like was seen when x264 algorithms got ported into x265.
If this QP + arbitrary options toggling is true, then WTF I don't even. After all these years of proper methods being discussed and after their inclusion in AOM which should give them more expertise in this, they should know better.

I would start to suspect actual malice here (marketing/PR deciding the results here?). Because they are really jumping through hoops here to not use the encoders in a way they are supposed with that QP bullshit.
mandarinka is offline   Reply With Quote
Old 14th December 2018, 19:05   #1119  |  Link
mandarinka
Registered User
 
mandarinka's Avatar
 
Join Date: Jan 2007
Posts: 705
Quote:
Originally Posted by benwaggoner View Post
With fixed QP and no adaptive quant? I can do a LOT better than that with some tuning
I think the main point is that the default settings which they made effort to override could do much better. CQP is inferior to CRF mode, that is common knowledge, one of the first things that a non-100%clueless x264/x265 user picks up. It has been established 10 years ago.
mandarinka is offline   Reply With Quote
Old 14th December 2018, 19:50   #1120  |  Link
Beelzebubu
Registered User
 
Join Date: Feb 2003
Location: New York, NY (USA)
Posts: 47
Quote:
Originally Posted by benwaggoner View Post
Is aq-mode=0?
I missed this one. According to the figure in the blog post, as well as the paper, the default encoder setting for aq-mode was used for x264/5. The only encoder where aq-mode was specifically disabled is vpxenc, probably because enabling AQ in libvpx causes BDRATE losses in both PSNR as well as VMAF.
Beelzebubu is offline   Reply With Quote
Reply

Tags
google, ngov, vp8, vp9, vpx, webm

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 05:57.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.