or Connect
AVS › AVS Forum › Display Devices › Display Calibration › MadVR - ArgyllCMS
New Posts  All Forums:Forum Nav:

MadVR - ArgyllCMS - Page 28

post #811 of 1681
Quote:
Originally Posted by Kukulcan View Post

Thanks zoyd! Very interesting the fact that you had those same impressions and that finally you've changed your mind, I'll take it as a suggestion.

Actually I ended up in a Excel calculator posted somewhere in a thread. I played with it and I'm realizing that with BT1886 there's a strong relationship between contrast and the rate at which gamma goes up. So displays with very high contrast may show lower black levels (compared to pure power law) even at 10 IRE, or even 5 for best ones.

I will make other test in the following days, especially with my vpr Planar, I suspect that with a good display/projector, BT 1886 becomes very desirable.

Probably it's better getting rid of bad displays rather than BT1886....
Hey I'd buy a better monitor, but I blew all my money on a calibrator! :P (Just kidding, though getting a new monitor legitimately isn't an option for me right now)
post #812 of 1681
Yesterday night I gave a careful look at my Planar vpr with a LUT loaded, and it confirmed me again what I've already been saying... the problem is bad displays and not BT1886. I also changed my mind and now I agree with zoyd.

With my vpr the situation is upside down! BT1886 gives more contrast than power law, bringing at the same time more shadow details. The default state of the vpr is a perfect 2.22 gamma curve, so it was very easy to make a comparison.

Yes I know, good stuff isn't cheap and we all would like to get the most we can from our video chain, but unfortunately this LCD technology dominance today is really a curse for video quality. They're simply not up to the task! They were born as business display and are perfect for that task, but be damned the day they have become the consumer standard for video material!
post #813 of 1681

The closest monitor we can buy that matches a CRT (but HD) is a Plasma.

 

Pros:

 

Excellent Contrast Ratio.

Excellent Viewing angle.

Excellent Colors.

Perfect Uniformity.

 

Cons:

 

Higher Energy consumption.

Minimum of 50" size (less is not 1080p).

Highly reflective screen (needs very dark room).

Burn-in (not in latest models).

post #814 of 1681
In Cons: you missed ABL.
That makes plasmas next to useless for accurate image display.

(Ok - I am from a professional background, so I am being very critical, but I really find ABL totally changes the image intent, and makes accurate calibration all but impossible)...

Steve
post #815 of 1681
Hi Steve
ABL... I don't know very well plasmas, but what you said reminds me to dynamic iris issue... if so, I would put this ABL at the top of drawback list with regard to calibration/characterization.

Let's hope that OLED displays become available for consumers. They should be a quantum leap forward, especially for black level and even more important, gamut at low levels!
post #816 of 1681
Quote:
...and even more important, gamut at low levels!

 

Why is wide gamut at low levels is important?

How can it improve the picture or the way our eyes see it?

How is it relevant to film making?

post #817 of 1681
I hope it's ok if I post this here. I'm planning to release a much-needed update to dispcalGUI asap (probably within the next few days if no major issues emerge). If you'd like to test it, a snapshot of the current development version is available here: http://sourceforge.net/projects/dispcalgui/files/snapshot/dispcalGUI-1.4.7.3.zip/download
In a nutshell, it has madTGP as well as madVR 3D LUT creation support, along with most of the other new options introduced in Argyll CMS 1.6.x.
See the ReadMe for a full changelog (it's quite large this time around due to there not having been a release in the past six months).

The workflow on page 1 of this thread can be re-created in dispcalGUI in the following way if you're interested:
  1. Follow A 1.-4. like outlined on page 1 of this thread.
  2. Start madTGP.exe
    • Enable "use fullscreen"
    • Enable "disable VideoLuts"
    • Enable "disable 3dlut"
  3. Start dispcalGUI.exe. It will ask for the Argyll CMS tools location if they are not found.
    • Choose "sRGB" under "Settings" (this will also reset video card gamma table to linear).
    • Select "madVR" under "Display device".
    • Remove the checkmark from the interactive display adjustment checkbox.
    • Select "XYZ LUT + Matrix" under "Profile type".
    • Select the "Massive" testchart under profiling settings.
    (Note: I could make this even easier by providing a "madVR" preset that could do all the above in one go. I'm just not sure about some parameters, e.g. should I really disable interactive display adjustment? It's easy enough to skip. Let me know what you think)
  4. Click "Calibrate & profile".
    After the calibration and profiling are done, you are given the option to install the profile. If you want to use it in color-managed apps, you may choose to do so, otherwise just click "Cancel".
  5. In the "Tools" menu, select "Create 3D LUT...".
    • Check the following:
      "Source profile" should already be filled with "Rec709.icm". Otherwise click "Choose..." to browse for it.
      "Destination profile" should already be filled with the display profile you just created. Otherwise click "Choose..." to browse for it, or "Current profile" to use the profile currently selected in dispcalGUI's main window under "Settings" (if that's "<Current>", the display profile assigned in Windows color management settings will be used, if any).
      "Apply calibration (vcgt)" should already be checked, otherwise, place a checkmark in the box.
    • Select "Absolute colorimetric with white point scaling" under "Rendering intent".
    • Select "madVR" under "3D LUT file format". This will automatically select the correct input/output encoding and 3D LUT size.
    • If you wish to use BT.1886 gamma mapping, place a check in the respective box. The setting equivalent to the commandline on page 1 of this thread is gamma 2.2, relative (this is the default).
    • Click "Create 3D LUT". Choose a location where to save the files. Wait. Done.
  6. Continue with section C on page 1 of this thread.
post #818 of 1681
Thread Starter 
Quote:
Originally Posted by fhoech View Post

I hope it's ok if I post this here. I'm planning to release a much-needed update to dispcalGUI asap (probably within the next few days if no major issues emerge). If you'd like to test it, a snapshot of the current development version is available here: http://sourceforge.net/projects/dispcalgui/files/snapshot/dispcalGUI-1.4.7.3.zip/download
In a nutshell, it has madTGP as well as madVR 3D LUT creation support, along with most of the other new options introduced in Argyll CMS 1.6.x.
See the ReadMe for a full changelog (it's quite large this time around due to there not having been a release in the past six months).

Thanks Florian! I will test this when I get a chance.

I'm currently troubleshooting color quality issues with the MadVR 3DLUTs generated. No matter which intent I use, most skin tones are looking like they are overly sunburned. Bright orange and deep magenta colored skin tones. I don't remember seeing this issue with 3DLUTs created 6 months ago and am not sure if there is a regression somewhere. I have been trying almost EVERY possible settings combinations. Different pattern sets, include/exclude calibration curves, different intents, madTPG or ArgyllCMS internal PG, etc.

Verifying the 3DLUTs with Calman do show < 2dE on 25% saturation sweeps, 25% luminance sweeps, and 24pt color checker patterns. Grayscale overall dE is less than 1.

Pulling my hair out on this!! confused.gif
post #819 of 1681
@fhoech,

sounds good to me, thanks! Unfortunately I don't have the time atm to test it.

FWIW, there's one tiny difference between the first page and my own recommended settings: I prefer to use "dispcal -qu" instead of "dispcal -qh". But I understand that using "-qu" can be painful if you have a slow meter. But at least with something as fast as the i1 Display 3 Pro, "dispcal -qu" is pretty quick, and I noticed a small improvement in quality. Just for what is worth...

Some sort of "intelligence" would be nice. E.g. if the user specifies that he wants to get a madVR compatible 3dlut, some other settings should be forced, or reduced. E.g. it would probably make no sense to use a simple matrix only profile in that situation. Not sure if the GUI could be optimized for this kind of logic? E.g. if the first combobox would be to ask the user what he wants to get in the end (e.g. a madVR 3dlut), that might allow you to reduce the options for the other combo/checkboxes, making life a bit easier for the user. But I'm not that familiar with dispcalGUI, so I'm not sure if that makes any sense, or how difficult it would be to implement.
post #820 of 1681
Quote:
Originally Posted by madshi View Post

Some sort of "intelligence" would be nice. E.g. if the user specifies that he wants to get a madVR compatible 3dlut, some other settings should be forced, or reduced. E.g. it would probably make no sense to use a simple matrix only profile in that situation. Not sure if the GUI could be optimized for this kind of logic? E.g. if the first combobox would be to ask the user what he wants to get in the end (e.g. a madVR 3dlut), that might allow you to reduce the options for the other combo/checkboxes, making life a bit easier for the user. But I'm not that familiar with dispcalGUI, so I'm not sure if that makes any sense, or how difficult it would be to implement.
That's along the line of thought I have with a (yet to be created) "madVR" preset. Basically this would set everything up with some good defaults. Currently for the actual 3D LUT creation some options are already forced when "madVR" is selected as format. I don't necessarily agree with ruling out matrix profiles completely though. If a display is behaving reasonably additive, you may actually get results that are "smoother" than with a LUT-type profile. If that's not the case a LUT profile may be required though, so I'm ok with that as a default.
post #821 of 1681
I had thought that a matrix profile wouldn't work if you want to get a 3dlut as the final output. If it does work then it's a valid option to offer, of course. In any case, a "madVR 3dlut" preset with good defaults and with all options limited to what makes sense would be a very nice solution.
post #822 of 1681
Thread Starter 
@fhoech

I'm getting a "dispcal.exe has stopped working" Windows error when I try to run a calibration using MadVR TPG. No issues using ArgyllCMS internal TPG.

Server 2012
ArgyllCMS v1.6.1 x86
DispcalGUI 1.4.7.3
i1D3
Edited by N3W813 - 10/10/13 at 1:22pm
post #823 of 1681
Quote:
Originally Posted by madshi View Post

I had thought that a matrix profile wouldn't work if you want to get a 3dlut as the final output. If it does work then it's a valid option to offer, of course. In any case, a "madVR 3dlut" preset with good defaults and with all options limited to what makes sense would be a very nice solution.
Matrix type profiles work fine for creating 3DLUTs - I'm using one right now as using a LUT type profile leads to awful black levels for my display (presumably the constraints on matrix + shaper curves force it to generate a more visually reasonable fit).
post #824 of 1681
Quote:
Originally Posted by N3W813 View Post

@fhoech

I'm getting a "dispcal.exe has stopped working" Windows error when I try to run a calibration using MadVR TPG. No issues using ArgyllCMS internal TPG.

Ah yes this is something I forgot, as it's not in the original instructions: You have to copy madHcNet32.dll from your madVR folder to where dispcal/dispread are (e.g. copy it to Argyll_V1.6.1\bin). It should work then.
post #825 of 1681
Thread Starter 
Hmmm, I copied madHcNet32.dll to the bin folder but I am still getting the same error. confused.gif
post #826 of 1681
Is madTGP.exe running? Do you maybe have another Argyll version somewhere on the PATH that dispcalGUI picks up? Can you try pointing it to the right version specifically? (File menu, locate Argyll executables)
post #827 of 1681
Thread Starter 
Quote:
Originally Posted by fhoech View Post

Is madTGP.exe running? Do you maybe have another Argyll version somewhere on the PATH that dispcalGUI picks up? Can you try pointing it to the right version specifically? (File menu, locate Argyll executables)

Yes, madTPG is running. No, I do not have another version of ArgyllCMS on the machine and I've checked the PATH system variable. I've ran the "Locate ArgyllCMS executables..." command and tried copying ArgyllCMS to another hard drive. Same issue. frown.gif

EDIT: Running the dispcal command (dispcal.exe -v -dmadvr -Yp -qh -m -w 0.3127,0.3290 -G2.4 -f0 VideoLUT) from a command prompt also fails so it is not caused by DispcalGUI. Maybe it is Server 2012? I will try it on my Win7SP1 HTPC later tonight.
Edited by N3W813 - 10/10/13 at 4:17pm
post #828 of 1681
Does
Code:
dispcal.exe  -v  -dmadvr  -Yp  -qh  -m  -gs  -w 0.3127,0.3290  "dispcal"
from the command prompt work?
post #829 of 1681
Thread Starter 
Issue resolved. I was using an older version of MadVR. Once updated to the latest version then copied the newest madHcNet32.dll to the ArgyllCMS bin directory, everything worked. smile.gif

@fhoech
Can you add a checkbox/option to disable the wait for meter prompt? When using madVR in fullscreen mode, one cannot click the OK button to continue the calibration.

As a workaround, I added "-Yp" to the additional command line options menu.
post #830 of 1681
Quote:
Originally Posted by N3W813 View Post

Issue resolved. I was using an older version of MadVR. Once updated to the latest version then copied the newest madHcNet32.dll to the ArgyllCMS bin directory, everything worked. smile.gif

Ok cool smile.gif
Quote:
Originally Posted by N3W813 View Post

@fhoech
Can you add a checkbox/option to disable the wait for meter prompt? When using madVR in fullscreen mode, one cannot click the OK button to continue the calibration.

As a workaround, I added "-Yp" to the additional command line options menu.

Done, it will now automatically skip the prompt when the madVR TPG is used. I've also gone ahead and added the madVR preset I talked about. So under step 3 in my post above you just select the madVR preset under settings.
New snapshot: http://dispcalgui.hoech.net/download/snapshot/dispcalGUI-win32.zip (this will always redirect to the latest snapshot if there is one)
[ I'd edit my post above, but it won't let me ]
post #831 of 1681
Quote:
Originally Posted by madshi View Post

@fhoech,

I prefer to use "dispcal -qu" instead of "dispcal -qh". But I understand that using "-qu" can be painful if you have a slow meter. But at least with something as fast as the i1 Display 3 Pro, "dispcal -qu" is pretty quick, and I noticed a small improvement in quality. Just for what is worth...

me too! Please Florian, consider to add the option -qu, it gives me consinstently lower DeltaE with 2 different displays (a DLP vpr and a LCD monitor). Of course I use the fast iD3 meter, that's why I can use it, otherwise it takes too long...
post #832 of 1681
@fhoech
Thanks for updating dispcalGUI, makes the whole process way more comfortable. It would be great if you would consider adding an option to view the commandline the current settings produce. I'm aware that you can see it in the log, but as far as I've seen only after actually starting a calibration.
post #833 of 1681
Quote:
Originally Posted by Kukulcan View Post

me too! Please Florian, consider to add the option -qu, it gives me consinstently lower DeltaE with 2 different displays (a DLP vpr and a LCD monitor). Of course I use the fast iD3 meter, that's why I can use it, otherwise it takes too long...
I'll think about it. If you say consistently lower dE, what order of magnitude are we talking about? And is it certain to be because of dispcal -qu and not some other setting?
post #834 of 1681
Well... actually I didn't make what a statician would call a controlled experiment. Anyway being on the safe side, I'd say a 0.5 DeltaE on average. As soon as I used qu, I found all grey from 10-100 IRE dropping below 1 DeltaE, especially with my Planar which is very stable and behaves pretty well, while with qh I had some of them up to 2. So more precisely, it's some spikes in dE that go down, and it happened also with my LCD screen, but in general the lcd measures are much less consistent...

I'm reporting it also because madshi has the same impression...
post #835 of 1681
Quote:
Originally Posted by monvo View Post

@fhoech
It would be great if you would consider adding an option to view the commandline the current settings produce. I'm aware that you can see it in the log, but as far as I've seen only after actually starting a calibration.
I've added a "dry run" setting to the options menu. When enabled, all functionality that calls Argyll CMS executables will effectively do nothing, but still print the command line parameters to the log.
I've updated the snapshot. http://dispcalgui.hoech.net/download/snapshot/dispcalGUI-win32.zip
post #836 of 1681
I just did a calibration of my JVC RS40 projector using the new version of DispcalGUI
(the same happens when using N3W813s batch files from page1 of this thread).
I have installed the new 3dlut in Madvr.

When using MadvrTPG as Generator for HCFR I get Delta Luma values of
+5.5 %-1.5 %-9.4 %+0.2 %-2.0 %+1.4 %
for RGBYCM.

However playing the test patterns from the AVS disc through MPC (or JRiver) and Madvr (with the new 3dlut installed)
gives these values:
-47.8 %-52.0 %-42.5 %-49.3 %-46.5 %-42.4 %

Any idea what's going on?
Edited by sepp - 10/11/13 at 9:57am
post #837 of 1681
Thread Starter 
@fhoech
Is the 'Verify Profile' under the Tools menu suitable for use for verifying a 3DLUT through madTPG? I've noticed that when running the verify profile process, the program automatically disables the VideoLUTs and the 3dlut in madTPG.

What is the process to verify the quality of the 3dlut through DispcalGUI?
post #838 of 1681
Quote:
Originally Posted by N3W813 View Post

@fhoech
Is the 'Verify Profile' under the Tools menu suitable for use for verifying a 3DLUT through madTPG? I've noticed that when running the verify profile process, the program automatically disables the VideoLUTs and the 3dlut in madTPG.

What is the process to verify the quality of the 3dlut through DispcalGUI?
'Verify profile' can currently only be used to verify the profile that was used as destination in the 3D LUT creation. While this does allow to draw some conclusions re the expected accuracy of the resulting 3D LUT, I realize this is not ideal. I plan to add 3D LUT (and device link) verification, but this will come in a later version, not in the one I'm planning to release soon.
What you can do right now with 'verify profile' is to check how well the profile handles the reproduction of e.g. Rec. 709. To do this, select the profile used as destination during 3D LUT creation under 'settings', then after choosing 'verify profile' from the menu, select the Rec709.icm profile, then any testchart (you can use one of the two provided 'verify' testcharts, but those are rather small. For a more thorough test, create a bigger chart with the testchart editor using a custom amount of patches. E.g. with some white patches, some single channel patches, some grayscale patches and any number of iterative patches with optimized farthest point sampling).
post #839 of 1681
Thread Starter 
@fhoech
In the dispcalGUI.log file, during a collink.exe command, if a BT.1886 curve is selected, the log output will only log the BT.1886 calculated values from 0% to 80%. 85%, 90%, 95%, and 100% values are not being logged.

If possible, in addition to the dispcalgui.log file, keep a separate log for each calibrate/profile session in its working folder? (ie. madVR 2013-10-11 0.3127x 0.329y Rec. 1886 S XYZLUT+MTX.log)

Thanks smile.gif
post #840 of 1681
Quote:
Originally Posted by N3W813 View Post

@fhoech
In the dispcalGUI.log file, during a collink.exe command, if a BT.1886 curve is selected, the log output will only log the BT.1886 calculated values from 0% to 80%. 85%, 90%, 95%, and 100% values are not being logged.

Good catch. I wasn't even aware that this was happening. Fixed now.
Quote:
Originally Posted by N3W813 View Post

If possible, in addition to the dispcalgui.log file, keep a separate log for each calibrate/profile session in its working folder? (ie. madVR 2013-10-11 0.3127x 0.329y Rec. 1886 S XYZLUT+MTX.log)

Thanks smile.gif

That will also make my life easier when helping out users, because I no longer have to hunt through huge logfiles (well not that huge atleast smile.gif)

The new snapshot (same link) contains the two above changes.
http://dispcalgui.hoech.net/download/snapshot/dispcalGUI-win32.zip
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › MadVR - ArgyllCMS