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 > Hardware & Software > Software players

Reply
 
Thread Tools Search this Thread Display Modes
Old 5th January 2017, 22:21   #41881  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
to my eyes this not minor here i can clearly see 8 or more steps on the back and i won't be shocked that it comes from double expending.
maybe that's the case with normal on this film stuff i don't watch stuff like this usually and i have debanding disabled by default. i now at least one scene where debanding low is kind of ruin it.
huhn is offline   Reply With Quote
Old 5th January 2017, 22:30   #41882  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
This is a example where debanding or true 10 bit would be required



__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004
CruNcher is offline   Reply With Quote
Old 5th January 2017, 22:36   #41883  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
Quote:
Originally Posted by CruNcher View Post
This is a example where debanding or true 10 bit would be required
if debanding works you don't need 10 bit...

most likely bit rate starved or encoded by an "not so good encoder" well samsung i guess...

BTW. is that a huge pixel error on the camera?
huhn is offline   Reply With Quote
Old 5th January 2017, 22:50   #41884  |  Link
Neo-XP
Registered User
 
Neo-XP's Avatar
 
Join Date: Mar 2016
Location: Switzerland
Posts: 140
Maybe this example convince you more (off/low/medium/high) :



Open in original size of you won't see anything.

Banding is still very noticeable on the top-left side with the "low" setting and you can clearly see here the details lost on the bottom of the image in the trees with the "high" setting.

Last edited by Neo-XP; 5th January 2017 at 22:57.
Neo-XP is offline   Reply With Quote
Old 5th January 2017, 23:23   #41885  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
Quote:
Originally Posted by huhn View Post
if debanding works you don't need 10 bit...

most likely bit rate starved or encoded by an "not so good encoder" well samsung i guess...

BTW. is that a huge pixel error on the camera?
debanding cant fix it you always gonna perceive it you need true precision and internal 10 bit 4:2:2 recording is entering now the Prosumer space it's starting this year at prices around 2000$

https://www.youtube.com/watch?v=rUnpEXaI1S8

4K 30 FPS 2:2:2 10 Bit HEVC

And yeah that red pixel error in Samsungs Demo is strange could be sensor issue could be post processing issue i dunno what it is it is definitely no decoding issue.

Also their whole processing is interesting

The Encoder on the outside here is ATEMEs Titan which does pretty good

Atemes Encoder pretty much dominates most of the Demos, especially LG and Samsung.
__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004

Last edited by CruNcher; 6th January 2017 at 00:09.
CruNcher is offline   Reply With Quote
Old 6th January 2017, 01:41   #41886  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
it is possible to create a smooth gradation with hat sky.

the banding is so terrible it can't be fixed with debanding completely but it is still a mastering issue not a recording.

8 bit is more than enough to show this smooth.
huhn is offline   Reply With Quote
Old 6th January 2017, 02:34   #41887  |  Link
baii
Registered User
 
Join Date: Dec 2011
Posts: 180
Quote:
Originally Posted by huhn View Post
i need 2100 mb with my RX 480 without NGU with NGU high it is over 2500 and NGU very high it is over 3000.

but finally someone that tested it too thanks
I was running ngu mid or low +3dlut. (Can't run mid +3d lut)
Memory can be lowered by lowering one of the queue, probably render or present, can't remember.
baii is offline   Reply With Quote
Old 6th January 2017, 02:37   #41888  |  Link
aufkrawall
Registered User
 
Join Date: Dec 2011
Posts: 1,812
Minor banding can also be observed in some Astra HEVC 10 bit videos, deband with low strength works wonders against it.
aufkrawall is offline   Reply With Quote
Old 6th January 2017, 03:06   #41889  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
Quote:
Originally Posted by baii View Post
I was running ngu mid or low +3dlut. (Can't run mid +3d lut)
Memory can be lowered by lowering one of the queue, probably render or present, can't remember.
deinterlancing doesn't work properly with lowered GPU queue so this is no real fix.

and even if you run out of Vram you can still run madVR but performance get's bad but that doesn't mean it will go over 40 ms.
huhn is offline   Reply With Quote
Old 6th January 2017, 03:26   #41890  |  Link
Georgel
Visual Novel Dev.
 
Georgel's Avatar
 
Join Date: Nov 2015
Location: Bucharest
Posts: 200
Quote:
Originally Posted by leeperry View Post
mVR already forces 0.5ms
Didn't know that madVR can force a change on the windows timer

Quote:
Originally Posted by tahaa7 View Post
OK, thanks. But, the problem with Intel 7th gen mobile processors is that you have to make a choice of either getting a quad-core CPU with a weaker iGPU, or a dual-core CPU with better (Iris Plus) iGPU.

There are more things to consider here.

If you're building a computer solely for madVR and video watching, a quad core or even dual core might be enough, but iGPU won't do that much for madVR either way.

If you want to do other things, depends on what you want to do.


Quote:
Originally Posted by huhn View Post
for madVR get the iris pro.

most content can be hardware decoded. an dual core is fine for most content anyway.

madVR doesn't care about CPU core numbers.
This is totally true true for a madVR usage case scenario ^

MadVR doesn't use much GPU by itself.
Georgel is offline   Reply With Quote
Old 6th January 2017, 05:42   #41891  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
Quote:
Originally Posted by aufkrawall View Post
Minor banding can also be observed in some Astra HEVC 10 bit videos, deband with low strength works wonders against it.
Do you see that Banding on a Premium UHDTV ?

Do you have a Native 10 bit Panel or 8 bit + FRC ?

Anyway Debanding is not a solution but a workaround and im pretty shocked that some find debanding the whole frames efficient same they like to enable dithering (i hate both and want it to finally die)

@ huhn
Of course you can do it better but that source is ok , anyway all these 10 bit demos will come out the most efficient on a real 10 bit Hardware chain that's fact and you know it.

You can throw Debanding and dithering over it and Destroy the result but that will stay a workaround nothing else that can be percepted (and has many issues) on a 55 UHDTV immediately your debanding and dithering will fail in such high gradiation recordings at that output Size especially in Motion you never get that transitiion band free 10 bit wasn't introduced for nothing and consumer can jump around that it's finally here.
__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004

Last edited by CruNcher; 6th January 2017 at 06:04.
CruNcher is offline   Reply With Quote
Old 6th January 2017, 05:50   #41892  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
If madVR's debanding can remove it the banding is probably from the source, not the display.
__________________
madVR options explained
Asmodian is offline   Reply With Quote
Old 6th January 2017, 06:08   #41893  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
Okay then lets compare hope you guys are also correctly calibrated to bt.709 in that case

This is what all UHDTVs in the middle class reach by now mostly perfectly and it has some impact on the final result if you wrong calibrated you most likely would see less banding because you gonna percept less details especially when your luminance is off and we talking about HDR stops here that will be recorded in the future by AVG Joe.

So lets see Spatialy

8 bit out no debanding



8 bit out debanding low (spatial/temporal)



Bands doesn't get completely eradicated looks still bad as before just some noise added which pushed up the filesize hurray

8 bit out debanding medium (spatial/temporal)



Filesize obviously gets higher it's bruteforce time

8 bit out debanding high (spatial/temporal)



Bruteforcing moves back a bit
__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004

Last edited by CruNcher; 6th January 2017 at 07:12.
CruNcher is offline   Reply With Quote
Old 6th January 2017, 06:58   #41894  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
It looks like there is bad banding in the source?

debanding is for smoothing out banding present in the source, not preventing banding due to low bitdepth output. New TVs might also have debanding capabilities, it is a very common problem.
__________________
madVR options explained
Asmodian is offline   Reply With Quote
Old 6th January 2017, 07:18   #41895  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
Of course its a very common problem because their is no real 10 bit chain working but you need that debanding wont help, the demo was exactly done to show that and Samsungs Native Gamut Range.
__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004

Last edited by CruNcher; 6th January 2017 at 07:33.
CruNcher is offline   Reply With Quote
Old 6th January 2017, 07:33   #41896  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
ok i have UHD TV with 12 bit input support and madVR supports a full 10 bit chain and i'm not going to waste my time on this because the issue is the source.

and to bad about BT 709 i'm currently using BT 1886.
huhn is offline   Reply With Quote
Old 6th January 2017, 07:40   #41897  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
Would be nice if you could confirm it is indeed the source which UHDTV is it exactly ?

http://demo-uhd3d.com/fiche.php?cat=uhd&id=91
__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004

Last edited by CruNcher; 6th January 2017 at 07:45.
CruNcher is offline   Reply With Quote
Old 6th January 2017, 07:44   #41898  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
i don't have the file and my TV can show a clear smooth gradation so it is clearly the file.

the file is bitrate starved, bad mastered or both.
huhn is offline   Reply With Quote
Old 6th January 2017, 08:03   #41899  |  Link
CruNcher
Registered User
 
CruNcher's Avatar
 
Join Date: Apr 2002
Location: Germany
Posts: 4,926
Phantom Flex has also lot of scenes that can band extremely and MadVR has no chance avoiding it

http://demo-uhd3d.com/fiche.php?cat=uhd&id=98

This footage is also partly used in "World in HDR" and SES Astra Demo



Same for Powerofcurve

http://demo-uhd3d.com/fiche.php?cat=uhd&id=148

It's also very visible that the Computer Generated stuff allways comes out better then the almost untouched natural light shoots




Ride on Board banding example

http://demo-uhd3d.com/fiche.php?cat=uhd&id=97










So it would be really nice if you could confirm that what i see is what you see on your 10 bit chain or to disconfirm because it makes no sense to show this 10 bit progress to the public if the sources are all wrong

Everyone would realize that on such big screens with a bad 10 bit chain and calibrated.

And im Pretty sure Ateme knows what they doing when it's about 10 bit
__________________
all my compares are riddles so please try to decipher them yourselves :)

It is about Time

Join the Revolution NOW before it is to Late !

http://forum.doom9.org/showthread.php?t=168004

Last edited by CruNcher; 6th January 2017 at 09:11.
CruNcher is offline   Reply With Quote
Old 6th January 2017, 08:13   #41900  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
Quote:
Originally Posted by huhn View Post
i don't have the file and my TV can show a clear smooth gradation so it is clearly the file.

the file is bitrate starved, bad mastered or both.
or you are not using madVR...:
https://abload.de/img/madvrldp0v.png

this is without debanding. i don't even want to know how you get so much banding in picture.
huhn is offline   Reply With Quote
Reply

Tags
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling

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 23:11.


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