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 > VP9 and AV1

Reply
 
Thread Tools Search this Thread Display Modes
Old 18th April 2018, 10:34   #641  |  Link
user822
Registered User
 
Join Date: Mar 2009
Posts: 54
I would say that the film grain synthesis is something that no one has really looked at yet in tests, that will be a huge advantage over HEVC.
Screenshots below are from
  1. Bluray Source (125MB H264)
  2. aom-master 2pass cq-level=40 (cpu-used=2 film-grain-test=1 bit-depth=10) (8.3MB)
  3. x265-master 2pass bitrate=2025 (preset=placebo output-depth=10) (8.0MB)
Frame 1:
Src, x265, aom
Frame 2:
Src, x265, aom
user822 is offline   Reply With Quote
Old 18th April 2018, 13:04   #642  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Quote:
Originally Posted by user822 View Post
I would say that the film grain synthesis is something that no one has really looked at yet in tests, that will be a huge advantage over HEVC.
Screenshots below are from
  1. Bluray Source (125MB H264)
  2. aom-master 2pass cq-level=40 (cpu-used=2 film-grain-test=1 bit-depth=10) (8.3MB)
  3. x265-master 2pass bitrate=2025 (preset=placebo output-depth=10) (8.0MB)
Frame 1:
Src, x265, aom
Frame 2:
Src, x265, aom
I still think it's the worst idea ever. Time would have been better spent creating an extraordinary grain filter that leaves useful detail intact. (as much as realistically possible)

Out of curiosity, can the synthesis part (but not the removal.) be disabled (or defeated) to see how it looks without it?
mzso is offline   Reply With Quote
Old 18th April 2018, 13:22   #643  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,565
Quote:
Originally Posted by mzso View Post
I still think it's the worst idea ever. Time would have been better spent creating an extraordinary grain filter that leaves useful detail intact.
That can be done purely as an encoder-side feature and it's not like degraining filters are something new. No need to have something done for the bitstream freeze.
sneaker_ger is offline   Reply With Quote
Old 18th April 2018, 13:39   #644  |  Link
MoSal
Registered User
 
Join Date: Jun 2013
Posts: 94
Quote:
Originally Posted by user822 View Post
I would say that the film grain synthesis is something that no one has really looked at yet in tests, that will be a huge advantage over HEVC.
If content providers use it.

I definitely look for the day where I no longer need to use the sharpening filter shortcuts I added to my player configuration.
__________________
saldl: a command-line downloader optimized for speed and early preview.
MoSal is offline   Reply With Quote
Old 18th April 2018, 14:32   #645  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Quote:
Originally Posted by MoSal View Post
If content providers use it.

I definitely look for the day where I no longer need to use the sharpening filter shortcuts I added to my player configuration.
How is that? I only feel the need to disable sharpening if the video is atrociously grainy/noisy. This keep-grain feature doesn't help with this at all. It might make it worse though.
mzso is offline   Reply With Quote
Old 18th April 2018, 16:21   #646  |  Link
MoSal
Registered User
 
Join Date: Jun 2013
Posts: 94
Quote:
Originally Posted by mzso View Post
How is that? I only feel the need to disable sharpening if the video is atrociously grainy/noisy. This keep-grain feature doesn't help with this at all. It might make it worse though.

I don't disable sharpening. I enable sharpening as I can't stand the smoothness of many videos.

I basically have this in ~/.mpv/input.conf:

Code:
k add sharpen +0.25
K add sharpen -0.25
Values between 0.75 and 1.25 look good for my eyes.

And I already use ewa_lanczossharp for scaling.
__________________
saldl: a command-line downloader optimized for speed and early preview.
MoSal is offline   Reply With Quote
Old 19th April 2018, 13:56   #647  |  Link
Barough
Registered User
 
Barough's Avatar
 
Join Date: Feb 2007
Location: Sweden
Posts: 481
AOM AV1 v0.1.0-9264-gbce84eb0b
Built on April 19, 2018, GCC 7.3.0

Code:
https://aomedia.googlesource.com/aom
Barough is offline   Reply With Quote
Old 20th April 2018, 02:35   #648  |  Link
foxyshadis
ангел смерти
 
foxyshadis's Avatar
 
Join Date: Nov 2004
Location: Lost
Posts: 9,556
HM and JM both had a rudimentary implementation of FGM, and they never made it into any of the commercial or major free encoders or decoders. I hope AV1's makes it into the wild, because that's one of the features I most sorely miss for mid-to-low bitrate encoding. There wouldn't be such a thing as "too smooth" at normal bitrates anymore; HEVC has really suffered among video enthusiasts because of that.
foxyshadis is offline   Reply With Quote
Old 20th April 2018, 13:32   #649  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Quote:
Originally Posted by ianken View Post
With a current FFMPEG build I get 0.2 fps on 1080p content. And that's a single stream.

To use this for next-day TV VOD it needs to be...faster. Or, maybe the ffmpeg integration is not fully baked? The CPU certainly is not loaded in any appreciable way. It's not even clocking up.
Oh, cool, at least we won't need to wait for it to get into ffmpeg. I missed this comment before. Although I think I'll hold back on trying it until some performance improvements land.

(BTW, It's been mentioned a bunch of times that they didn't do any performance optimizations in libaom.)
mzso is offline   Reply With Quote
Old 20th April 2018, 18:08   #650  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 6,746
AOM v0.1.0-9271-gd6499b0cd
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 20th April 2018, 18:32   #651  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Quote:
Originally Posted by Tommy Carrot View Post
The bitstream is not frozen, so you should use av1 only for testing... but currently it's too slow even for that. Once it is frozen, i expect the speed optimizations will come fairly quickly. Their stated goal is around a hundred times faster encoding speed at the end of the year. Hopefully with not significant quality drop.
Well, at least it maximizes it's single-thread cpu utiliztation. I get constant 8.32-8.33 CPU utilization on my 12 (virtual) core system.

Last edited by mzso; 20th April 2018 at 18:40.
mzso is offline   Reply With Quote
Old 20th April 2018, 21:51   #652  |  Link
Blue_MiSfit
Derek Prestegard IRL
 
Blue_MiSfit's Avatar
 
Join Date: Nov 2003
Location: Los Angeles
Posts: 5,986
^^ Is it just me, or does that sound extremely optimistic?

The bitstream was supposed to be frozen about 18 months ago I think? Maybe they should be a bit more conservative with timelines.

Last edited by Blue_MiSfit; 20th April 2018 at 21:54.
Blue_MiSfit is offline   Reply With Quote
Old 22nd April 2018, 08:23   #653  |  Link
WhatZit
Registered User
 
Join Date: Aug 2016
Posts: 60
Quote:
Originally Posted by Blue_MiSfit View Post
Maybe they should be a bit more conservative with timelines.
Remember that the fundamental purpose behind AV1's existence is to spit in the eye of HEVC.

Should the AoM have published credible timelines, HEVC would have had ample opportunity to dodge out of the way of any expectoration. Whether they'd have actually taken advantage of those opportunities is another matter.

As it currently stands, the decision facing the streaming industry is this: with the rapid uptake of UHD/HDR consumer displays (aka immediate demand), do you implement a UHD delivery system NOW using HEVC's absurd licensing, or in 3 years time using AV1's open license?

That decision should be a no-brainer... if it weren't for the likes of Velos being in the HEVC patent pool.

So, if the AoM were able to convince those decision makers that they need only wait 1 year (irrespective of the timeline's veracity), HEVC gets it in the eye again.
WhatZit is offline   Reply With Quote
Old 22nd April 2018, 20:45   #654  |  Link
iwod
Registered User
 
Join Date: Apr 2002
Posts: 756
Quote:
Originally Posted by Blue_MiSfit View Post
^^ Is it just me, or does that sound extremely optimistic?

The bitstream was supposed to be frozen about 18 months ago I think? Maybe they should be a bit more conservative with timelines.
Also remember they are still pretty much the guys from On2, ( Or at least half of them ) anyone who has been in Doom9 long enough would know what they are like.
iwod is offline   Reply With Quote
Old 24th April 2018, 15:59   #655  |  Link
foxyshadis
ангел смерти
 
foxyshadis's Avatar
 
Join Date: Nov 2004
Location: Lost
Posts: 9,556
Quote:
Originally Posted by iwod View Post
Also remember they are still pretty much the guys from On2, ( Or at least half of them ) anyone who has been in Doom9 long enough would know what they are like.
I don't think there's that many On2 employees left; I think it's more a general quality of the AV industry to overpromise and underdeliver, and Google itself is well-known for unfounded enthusiasm for eternal betas. Kind of a match made in heaven, there.
foxyshadis is offline   Reply With Quote
Old 24th April 2018, 20:20   #656  |  Link
Clare
Registered User
 
Join Date: Apr 2016
Posts: 61
Added AV1 to my comparison: https://wyohknott.github.io/video-formats-comparison/

I had to run it at cpu-used=4 to get a reasonable encode time so it's not representative of the max quality that could be obtained.

You can see that the encode speed is still infinitesimal.
Clare is offline   Reply With Quote
Old 25th April 2018, 07:39   #657  |  Link
Shevach
Video compressionist
 
Join Date: Jun 2009
Location: Israel
Posts: 126
Quote:
Originally Posted by foxyshadis View Post
Doesn't seem to have ever been discussed in public. I bet it was a test an engineer inserted and it was just never removed; I wonder if it's ever been used in real life.
Rationales of many AV1 features seem me vague. Woo-Jin Han in LinkedIn forum "AV1 Learning" commented 1/8-pel precision:
" It was proven that high precision mc is only effective for low resolution. Beyond 1080p, even 1/4 pel seems not very effective to justify its complexity. I remember that 1/4 pel shows 30% coding gain for famous Mobile CIF sequence (low res, complex texture, slow motion, aliasing from inadequate down sampler) but average 5-8% over various test sequences. Very high frame rate with non hierarchical structure may change the situation but it seems clear that it is highly sequence dependent tool"

I remember a paper "Motion-Compensating Prediction with Fractional-Pel Accuracy", by B. Girod, 1993 . If we omit a mathematical part of the article and go to the conclusion part, it's written (in my wording): for blocks 16x16 of TV video resolution 1/4-pel motion accuracy appears to be sufficient.
Shevach is offline   Reply With Quote
Old 25th April 2018, 07:46   #658  |  Link
Shevach
Video compressionist
 
Join Date: Jun 2009
Location: Israel
Posts: 126
AV1 enables to omit transmission of skip_flags - what's a gain in coding efficiency?

AV1 enables to omit transmission of skip flags. Indeed, there is the frame-header parameter 'skip_mode_present' and if this parameter equals to 1 then the skip flags are not signalled. The rationale is clear, if we know 'a priory' that encoding of a given frame will not produce skip blocks then it's redundant to transmit skip flags.
Let's look at the situation from another view, what's a penalty in transmission of skip flags provided that all blocks are non-skipped?
Because i am not completely familiar with AV1 entropy coding process i consider AVC/HEVC arithmetic engine instead.
In HEVC/AVC the maximal probability of a symbol is ~0.98 and hence the number of bits produced by encoding a symbol having the maximal probability is -log2(0.98) = ~0.02 bits.
Let's suppose that we know ahead that all MBs will be no-skipped (i.e. skip_flag = 0 for each MB) and consequently the skip_flag syntax element gets maximal probability 0.98. In such case the total number of bits consumed by all skip flags is ~0.02 x Number_Mbs.
For example, HD resolution frame has usually 8100 MBs (16x16 grid) and hence the total size of all skip_flag syntax elements is ~8100 x 0.02 = 162 bits. It's negligible in most cases.
Probably AV1 frame level parameter 'skip_mode_present' (to disable transmission of skip flags) has a minimal impact on coding efficiency, although it increases the decoding complexity (more 'if-else').
Shevach is offline   Reply With Quote
Old 25th April 2018, 15:42   #659  |  Link
Shevach
Video compressionist
 
Join Date: Jun 2009
Location: Israel
Posts: 126
Quote:
Originally Posted by Clare View Post
Added AV1 to my comparison: https://wyohknott.github.io/video-formats-comparison/

I had to run it at cpu-used=4 to get a reasonable encode time so it's not representative of the max quality that could be obtained.

You can see that the encode speed is still infinitesimal.
What's sense to check codecs with PSNR values 45 dB and greater?
According to HVS research all video with PSNR above 45 dB look perceptual identical to the original.
Shevach is offline   Reply With Quote
Old 27th April 2018, 07:12   #660  |  Link
Djfe
Registered User
 
Join Date: May 2014
Posts: 24
Quote:
Originally Posted by mzso View Post
I still think it's the worst idea ever. Time would have been better spent creating an extraordinary grain filter that leaves useful detail intact. (as much as realistically possible)

Out of curiosity, can the synthesis part (but not the removal.) be disabled (or defeated) to see how it looks without it?
Use the NL-Means filter on Handbrake instead (way too slow for an encoder but beautiful/very good IMO)


Another topic:
Good video to explain people what av1 is, why it was created and by whom etc. (the whole story):
https://youtu.be/lEdqN22vaWs
Djfe is offline   Reply With Quote
Reply

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 04:30.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.