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 4th February 2018, 20:19   #48821  |  Link
SamuriHL
Registered User
 
SamuriHL's Avatar
 
Join Date: May 2004
Posts: 4,267
Quote:
Originally Posted by ryrynz View Post
Yeah SSIM 2D is sharper though with less artifacts and I don't want to settle.
I totally understand that one. My HTPC is probably going to undergo a windows reinstall soon so I can get it back to working status. 3D has been a problem but then I was just playing some recorded TV and it locked up my machine twice. That's not what I call acceptable. LOL. So I may get to play with a clean install and see if I can get SSIM2D working better.
__________________
HTPC: Windows 10, I9 9900k, RTX 2070 Founder's Edition, Pioneer Elite VSX-LX303, LG C8 65" OLED
SamuriHL is offline   Reply With Quote
Old 4th February 2018, 20:26   #48822  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by ryrynz View Post
Cool. Any quick info on the hue changes, say comparative quality to the HQ setting and performance?
The new "do *not* preserve hue" option should (hopefully) combine the best of the previous "do not preserve hue" and "preserve hue with low quality" options. And it might be a bit faster than before, too.

The new "preserve hue" option is almost the same as the old "high quality" version, but with a saturation bugfix. Speed probably no noticeable change.

Quote:
Originally Posted by thighhighs View Post
Thx for new build! Any info about upcoming features?
Nope...

Quote:
Originally Posted by Neo-XP View Post
Any explanation of what it does exactly please?

What are the differences between all these?

- 100 Nits
- 50 Nits, linear
- 50 Nits, spline
- 50 Nits, linear, desaturate
- 50 Nits, spline, desaturate
I've been told that projector owners in a bat cave "officially" target diffuse/specular white at 50 Nits instead of 100 Nits. The HDR transfer function defines for each pixel with which Nits it should be displayed. But that's for 100 Nits diffuse white. So the new option practically scales all HDR pixels down by a factor of 2. So a pixel which should be drawn at e.g. 40 Nits is now drawn at 20 Nits instead.

I don't really know yet what practical effect that will bring. I'll wait for projector users to report back if the new setting is helpful at all or not.

The "linear" option simply halfs each pixel's brightness. The "spline" option tries to keep the original brightness in the shadow detail area, while still ending up at half brightness at diffuse white. I'm not really happy with the spline option yet, the curve probably needs some work.

The "desaturate" option does processing slightly different which results in ever so slightly lower saturation. I'm not sure which is "better" or more correct. So I'm offering both for testing. Of course one will go away once testing has been completed.
madshi is offline   Reply With Quote
Old 4th February 2018, 20:36   #48823  |  Link
Ver Greeneyes
Registered User
 
Join Date: May 2012
Posts: 445
Thanks for the new release!

Quote:
Originally Posted by madshi View Post
Code:
* HDR option "fix too bright & saturated pixels" is now always available
This is purely nitpicking of course, but I think this would flow better as "fix overly bright & saturated pixels".
Ver Greeneyes is offline   Reply With Quote
Old 4th February 2018, 20:51   #48824  |  Link
ABDO
Registered User
 
Join Date: Dec 2016
Posts: 64
thanks @madshi for new madvr build
ABDO is offline   Reply With Quote
Old 4th February 2018, 21:43   #48825  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 5,932
Code:
* added option "use reduce compression artifacts only if it comes for free"


Documented to work with NGU resampling; so I tried it with medium AA option. Looks quite good in a rather dark video. Not sure if really fluid, though. I'll test further from here on...

__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 4th February 2018, 21:51   #48826  |  Link
Manni
Registered User
 
Join Date: Jul 2014
Posts: 717
Thanks for the new build Madshi. I still don't have much time right now but I tried to take a quick look.

The new diffuse white option is very useful to us projector owners in a dedicated room, so thanks for implementing this. I didn't have the time to test in depth but it's promising and goes in the direction of what I was asking. You might want to give more options, as although the standard is 50nits for SDR, with HDR we need to go down to 16-20nits in some cases, so I would suggest 25-50 as possible values (factor of 2 or 4).

I still have issues with the way highlights are being processed with the pixel shader conversion, and now that I can rule out the 3D LUT (with the new line in the OSD, super useful so thanks for that), there is something definitely wrong with MadVR's processing.

I specified - as you suggested - my true peakY which is 125nits. I've tested with "This display is already calibrated" and with a null 3D LUT for SDR BT2020 generated with Lightspace (checked and confirmed as not changing anything to the picture).

In both cases, there is very significant posterization/clipping (depending on the settings) with the pixel shader conversion that entirely goes away in passthrough when using a custom curve in the projector. If you don't have a way to upload a custom curve, trust me and take a look at the scene below with the pixel shader conversion.

Quick question: in passthrough, the OSD reports that it's using the SDR-BT2020 3D LUT if 3D LUTs are selectin in the calibration tab. Is it to be expected? It should either use nothing (as it's sending HDR, not SDR) or use an HDR BT2020 3D LUT (I think).

Please try the beginning of Mad Max, minute 3:20. The end of the shot is a sunset with dust settling down and a fade out. The posterization/clipping (depending on settings) on the red sun/sky is very, very bad.

I thought it was my 3D LUT but it isn't, as even a null LUT (or "this display is already calibrated") produces the same effect.

Let me know by email/PM if you need a short clip of the beginning of Mad Max (as well as the null 3D LUT I'm using for these tests if you want to check it).

Any progress on the custom API to send the SDR BT2020 info in the HDMI stream?

EDIT: I have since this that you requested feedback in this thread on AVS, so I posted a copy of this post there, as it's probably a better place to discuss this.
EDIT2: Madshi has confirmed that this is being discussed in the thread linked above as there are more projector owners on AVS, so if you are interested please follow the discussion there.
__________________
Win10 Pro x64 b1809 MCE
i7 3770K@4.0Ghz 16Gb@2.18Ghz EVGA GTX 1080 Ti SC2 11Gb@2GHz 398.11 RGB Full 8bits
MPC-BE/LAV/MadVR/jRiver/MyMovies V5.25
Denon X8500H>HD Fury Maestro>JVC RS2000

Last edited by Manni; 4th February 2018 at 23:04.
Manni is offline   Reply With Quote
Old 4th February 2018, 21:56   #48827  |  Link
zapatista
Registered User
 
Join Date: Nov 2016
Posts: 16
Hi Madshi,

is there any possibility madvr would be able to downconvert video in 4k rec.2020 (or DCI-p3) color format to to 1080 resolution in x.v.Color (eg xvYCC, or Extended-gamut YCC)

there are a significant number of higher end 1080 pannels who can display x.v.Color (which is very close to the rec2020 color space), allowing those users to potentially benefit from the higher color detail/quality found in 4k UHD bluray (in addition to the 4k YUV 4.2.0 to 4k RGB 4.4.4 conversion step you have already implemented, before the 4k video is down converted to 1080 10 bit)

Last edited by zapatista; 4th February 2018 at 21:59.
zapatista is offline   Reply With Quote
Old 4th February 2018, 22:07   #48828  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 5,823
it's very unlikely that a display that support x.v color space can't display all these colors out side this mode so a 3D lut give you the same possibility.

just to make something every clear sony come up with this while 1080p sony TV are often used panels that can't reach 100 % BT 709.
and next problem can GPUs even handle this dead colorspace?
huhn is offline   Reply With Quote
Old 4th February 2018, 22:25   #48829  |  Link
YxP
Registered User
 
Join Date: Oct 2012
Posts: 97
* added option "use reduce compression artifacts only if it comes for free"

Laughed at this, feels like it was added for me and other stupids But thanks! (btw using RCA@1 almost all the time)
YxP is offline   Reply With Quote
Old 4th February 2018, 22:29   #48830  |  Link
YxP
Registered User
 
Join Date: Oct 2012
Posts: 97
Btw if it's any interest to you, the GPU sound is almost eliminated when using D3D11 decoder in lavfilters.
YxP is offline   Reply With Quote
Old 4th February 2018, 22:34   #48831  |  Link
Manni
Registered User
 
Join Date: Jul 2014
Posts: 717
Quote:
Originally Posted by zapatista View Post
Hi Madshi,

is there any possibility madvr would be able to downconvert video in 4k rec.2020 (or DCI-p3) color format to to 1080 resolution in x.v.Color (eg xvYCC, or Extended-gamut YCC)

there are a significant number of higher end 1080 pannels who can display x.v.Color (which is very close to the rec2020 color space), allowing those users to potentially benefit from the higher color detail/quality found in 4k UHD bluray (in addition to the 4k YUV 4.2.0 to 4k RGB 4.4.4 conversion step you have already implemented, before the 4k video is down converted to 1080 10 bit)
x.v.color is not very close to BT2020, it's an extension of rec-709 that has no standard (it varies from manufacturer to manufacturer). With some manufacturers, it's a joke. For example, many Sony projectors reach barely more than rec-709 in x.v.color because they don't have a DCI-P3 filter. JVC projectors with a P3 filter do better. In any case, it's not possible to calibrate to x.v.color because it doesn't have any coordinates (it's entirely relative to rec-709 and dependent on each display's implementation), so I highly doubt that Madshi could do something like this.

You should use the native gamut of the display and create a 3D LUT to BT2020 or DCI-P3 if your display doesn't have a BT2020 or P3 mode. What matters most isn't the saturation at 100%, it's the way saturation will track within the gamut. You can get very decent results even with a display not able to reach much more than rec-709 as long as the saturations are tracking BT-2020.
__________________
Win10 Pro x64 b1809 MCE
i7 3770K@4.0Ghz 16Gb@2.18Ghz EVGA GTX 1080 Ti SC2 11Gb@2GHz 398.11 RGB Full 8bits
MPC-BE/LAV/MadVR/jRiver/MyMovies V5.25
Denon X8500H>HD Fury Maestro>JVC RS2000
Manni is offline   Reply With Quote
Old 4th February 2018, 22:44   #48832  |  Link
MS-DOS
Registered User
 
Join Date: Sep 2012
Posts: 76
madshi, I don't recall if you answered this already, do you plan moving black bars detection to GPU?
MS-DOS is offline   Reply With Quote
Old 5th February 2018, 00:55   #48833  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 5,823
Quote:
Originally Posted by LigH View Post
Code:
* added option "use reduce compression artifacts only if it comes for free"


Documented to work with NGU resampling; so I tried it with medium AA option. Looks quite good in a rather dark video. Not sure if really fluid, though. I'll test further from here on...

it's never free with AA only with soft, standard and sharp.
maybe AA should be renamed so it is not so misleading and unlike the other NGU algorithm quite different.

and it's quite an "issue" for me because i'm generally favouring AA for bad sources that need RCA.
huhn is offline   Reply With Quote
Old 5th February 2018, 01:54   #48834  |  Link
zapatista
Registered User
 
Join Date: Nov 2016
Posts: 16
Quote:
Originally Posted by Manni View Post
I highly doubt that Madshi could do something like this.

You should use the native gamut of the display and create a 3D LUT to BT2020 or DCI-P3
thank you for the answer and the clear explanation (even if i dont yet understand it fully)

i should have been a little more precise in my original wording
- most current tv models with UHD displays using rec.2020 can still not display the full range of this color space ( or the full range of DCI-P3 ), and in their specs this is quantified in a %
- for ex a 2015 higher end samsung UHD tv (model 65JS9500) (using rec.2020) can still only display 91% coverage of DCI-P3 colour gamut
- in comparison, my 1080p sony tv panel (with HD using rec.709) is able to display 94.7% of the DCI-P3 colour space (using the x.v.color setting i presume), and it can accept this signal over HDMI input (and my HTPC can output x.v.color)
- DCI-3 gamut and the x.v.color gamut are VERY similar in size (but use slightly different primaries)


so if i can convert a 4k video (yuv 4.2.0 10 bit) to 4k RGB 4.4.4, and then in the last conversion step (resolution 4k to 2k) i am hoping to maintain as much of the original color information by converting the rec.2020 UHD color gamut % used by the original file, to x.v.color at the highest % my display can cope with (instead of squashing it into rec.709)

i dont really know what Lut's are or how i can use them ( ? look up tables for the RGB primaries used by the display ?), i was hoping the conversion from UHD video to HD (but in x.v.color instead of rec.709 ) could do this in madvr

is there an easy way i can find out my tv model x.v.color primaries, and then create a 3D lut file, and then load it in madvr ? or would it be possible to enter my information directly into madvr maybe
Attached Images
 
zapatista is offline   Reply With Quote
Old 5th February 2018, 02:29   #48835  |  Link
JarrettH
Registered User
 
Join Date: Aug 2004
Posts: 838
I read this in the other thread about madvr settings.

Quote:
run RCA as part of NGU, if possible: [Enabled] Combines most of the work required for 'reduce compression artifacts' into NGU Sharp upscaling for a large performance advantage when both are run at the same time.
With this option enabled medium quality forces NGU Sharp to high and high quality forces NGU Sharp to very high.
Why is there this behaviour? So it's only free if you can run NGU high or NGU very high? I run NGU luma medium, and yep it forces it to high, which I can't do.

Last edited by JarrettH; 5th February 2018 at 04:11.
JarrettH is offline   Reply With Quote
Old 5th February 2018, 02:58   #48836  |  Link
ryrynz
Registered User
 
ryrynz's Avatar
 
Join Date: Mar 2009
Posts: 3,203
That option has been removed now. I don't think that the requirements have changed.. Still NGU/NGU sharp high/very high.

I believe madshi was looking to add RCA for free with NGU AA it's just as huhn said quite different so requires extra code.

Last edited by ryrynz; 5th February 2018 at 03:22.
ryrynz is offline   Reply With Quote
Old 5th February 2018, 04:59   #48837  |  Link
har3inger
Registered User
 
Join Date: Feb 2014
Posts: 139
To clarify:

When RCA tries to run for free, will it still run on NGU'd chroma channel, if the luma channel is not scaled by NGU/already full res native? Does the checkbox "process chroma as well" do anything when the algo tries to run for free?
har3inger is offline   Reply With Quote
Old 5th February 2018, 07:12   #48838  |  Link
Sebastiii
Registered User
 
Join Date: Oct 2009
Location: France
Posts: 613
Quote:
Originally Posted by Clammerz View Post
Wrong thread. This is for MadVR. Not LAVFilters.
From madVR itself and happening on both different player.
But to be sure will switch on EVR to confirm it at 100%
__________________
HTPC : i7 920 6Go Win10(x64) / Nvidia 1050Ti / P6T Deluxe / Harman-Kardon AVR-355.
Sebastiii is offline   Reply With Quote
Old 5th February 2018, 07:15   #48839  |  Link
Sebastiii
Registered User
 
Join Date: Oct 2009
Location: France
Posts: 613
Confirmed it comes from madVR and the OSD message in left upper corner is the same font as when changing madVR option from keyboard shorcut, so yes @Clammerz i'm right to asking this here
__________________
HTPC : i7 920 6Go Win10(x64) / Nvidia 1050Ti / P6T Deluxe / Harman-Kardon AVR-355.
Sebastiii is offline   Reply With Quote
Old 5th February 2018, 07:20   #48840  |  Link
Sebastiii
Registered User
 
Join Date: Oct 2009
Location: France
Posts: 613
Quote:
Originally Posted by madshi View Post
Weird. You should get that message if you use an outdated LAV build (because some older versions produced crashes). But you should not get it with a recent nightly. What exact version number does your "LAVSplitter.ax" file have?
Hi
Thanks for the new build will try asap
So the version from LAVSplitter is on older one, so you are right, i have installed a tool this week-end that install on older version of LAV suite.
Sorry, that's shame that i didn't look into it myself.

LAVAudio and LAVVideo was the correct version but not LAVSplitter.
__________________
HTPC : i7 920 6Go Win10(x64) / Nvidia 1050Ti / P6T Deluxe / Harman-Kardon AVR-355.
Sebastiii 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 23:58.


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