MadVR - ArgyllCMS - Page 88 - AVS Forum
Forum Jump: 
 57Likes
Reply
 
Thread Tools
post #2611 of 2785 Old 09-10-2014, 10:25 AM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
Quote:
Originally Posted by zoyd View Post
Typically a user should know via brightness test patterns whether or not the first few input codes are properly stimulating the display. They won't know if the absolute levels are correct but it's standard practice to ensure that level 16 is black and level 17 is the first code that generates any above black response.
Hmm. Given a native output offset gamma 2.2 curve, the brightness difference from level 16 (black) to 17 would be around 0.0007291753 cd/m2 (assuming a white level of 120cd/m2) if I'm not mistaken. Will that be visible in typical test patterns?
On my own LCD TV (native response roughly output offset gamma 2.2), when displaying a fullscreen black window with a 1/4 square of one level above black in the center, I'm having a very hard time trying to discern the square from the black background even if I almost completely darken the room (only some faint light coming in from tiny slits where the blinds don't quite fit exactly). Moving the square around helps, but it's still hard to see. If I look at the screen from the side, it becomes much easier to see due to effects of the viewing angle.
fhoech is offline  
Sponsored Links
Advertisement
 
post #2612 of 2785 Old 09-10-2014, 11:22 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,725
Mentioned: 30 Post(s)
Tagged: 0 Thread(s)
Quoted: 258 Post(s)
Liked: 361
Perhaps it's more difficult to do with LCDs, with a plasma it's very easy to tell the control position which just fires some pixels on pattern 17. Also with elevated minimum luminance (mine is 0.05 cd/m^2) the near-black per step luminance you want to end up with is not what you calculate assuming 2.2 gamma at those levels but closer to 1.6 gamma or 0.02 cd/m^2 for 120 cd/m^2 peak white. It looks like you are starting out in a clipped condition unless your native black point is much lower than mine. So prior to running dispcal you should be able to adjust brightness to see a difference between 16 and 17 for displays with these black levels and above. Make sure you are completely dark adapted and it also helps to using a flashing bar pattern rather than a static one.

Last edited by zoyd; 09-10-2014 at 12:05 PM.
zoyd is online now  
post #2613 of 2785 Old 09-10-2014, 04:20 PM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,725
Mentioned: 30 Post(s)
Tagged: 0 Thread(s)
Quoted: 258 Post(s)
Liked: 361
So I ran a couple of tests and I'm confused about one issue. I purposely set my brightness control to clip up to level 22 and then ran:

dispcal -v -dmadvr -X d3.ccmx -qm -f0 -G2.4 -k0 cal

for two different level paths

a. madTPG 0-255, GPU 0-255, display 0-255
b. madTPG 16-235, GPU 0-255, display 16-235

In the cal file it appears both cases moved input 0 to the correct output code to align with display black but when testing only case a. worked properly using madVR and a black clipping pattern. I tested by loading the gamma ramps to the video card after madVR was running. case b showed a black level rise. So does that mean the GPU ramps need to be scaled in that scenario? Is this something that is automatically done when including the video card calibration in the 3dlut flow (via the -e/-E switches)?
zoyd is online now  
post #2614 of 2785 Old 09-10-2014, 11:16 PM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,477
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 64 Post(s)
Liked: 132
The 3dlut is always in video levels. All the output level scaling in madVR is done after 3dlut processing. So basically for 3dlut processing in madVR there's no difference between 0-255 or 16-235 (or even custom) output levels. When using madTPG/madVR the calibration software (either ArgyllCMS or HCFR) should not need to worry at all about output levels. It should be totally transparent to the calibration software. All the output level scaling should be done inside of madVR/madTPG in such a way that the calibration software doesn't even have to know about it.

Of course when loading gamma ramps into the GPU, the situation is different. I don't really know how to handle different output levels there.
madshi is offline  
post #2615 of 2785 Old 09-11-2014, 03:24 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,725
Mentioned: 30 Post(s)
Tagged: 0 Thread(s)
Quoted: 258 Post(s)
Liked: 361
Quote:
Originally Posted by madshi View Post
The 3dlut is always in video levels. All the output level scaling in madVR is done after 3dlut processing.
I understand that but I was looking more for affirmation that the gamma ramp calibration information (.cal file) is scaled to video levels when using either the -a or -H switch to include it in madVR processing. I assume it is but just thought I'd check.
zoyd is online now  
post #2616 of 2785 Old 09-11-2014, 09:16 AM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
Quote:
Originally Posted by zoyd View Post
Perhaps it's more difficult to do with LCDs, with a plasma it's very easy to tell the control position which just fires some pixels on pattern 17.
I think this depends on the individual display and its transfer characteristics.

Quote:
Originally Posted by zoyd View Post
Also with elevated minimum luminance (mine is 0.05 cd/m^2) the near-black per step luminance you want to end up with is not what you calculate assuming 2.2 gamma at those levels but closer to 1.6 gamma or 0.02 cd/m^2 for 120 cd/m^2 peak white.
True, although the near black per step luminance that I ultimately want depends also on the 3D LUT parameters (usually I'm targeting BT.1886-like but with effective gamma 2.2 at 50%, and I've started playing with some output offset instead of all-input offset since Graeme made this available in the Argyll beta development code), while in this case I'm talking purely about display adjustment pre-characterization/LUT. In my case the gamma setting of the TV that gives me a gamma closest to 2.2 at 50% (which I'm usually aiming at) results in a curve that is close to an output offset gamma 2.2 curve.

Quote:
Originally Posted by zoyd View Post
It looks like you are starting out in a clipped condition unless your native black point is much lower than mine. So prior to running dispcal you should be able to adjust brightness to see a difference between 16 and 17 for displays with these black levels and above. Make sure you are completely dark adapted and it also helps to using a flashing bar pattern rather than a static one.
It's definitely not clipped, otherwise I wouldn't be able to see a difference at all. Moving/flashing pattern does help, also looking at the screen at an angle due to being an LCD. My i1 DisplayPro tells me the black level is 0.032971 cd/m^2 (white 120.4).
fhoech is offline  
post #2617 of 2785 Old 09-11-2014, 10:19 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,725
Mentioned: 30 Post(s)
Tagged: 0 Thread(s)
Quoted: 258 Post(s)
Liked: 361
Quote:
Originally Posted by fhoech View Post

True, although the near black per step luminance that I ultimately want depends also on the 3D LUT parameters (usually I'm targeting BT.1886-like but with effective gamma 2.2 at 50%, and I've started playing with some output offset instead of all-input offset since Graeme made this available in the Argyll beta development code), while in this case I'm talking purely about display adjustment pre-characterization/LUT. In my case the gamma setting of the TV that gives me a gamma closest to 2.2 at 50% (which I'm usually aiming at) results in a curve that is close to an output offset gamma 2.2 curve.
I see what you are saying but the point is that there are cases in which manual alignment of the display black point to input code 0 achieves better results than what dispcal is producing. So how do you handle that situation? The user can reset brightness after dispcal to remove any measurable above black offset or hack the .cal file, but neither of these options seems satisfactory. The algorithm could be improved but's in unclear where the problem lies, maybe the algorithm is fine and quantization in the video chain or measurements limits the performance.
zoyd is online now  
post #2618 of 2785 Old 09-11-2014, 10:56 AM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
Quote:
Originally Posted by zoyd View Post
I see what you are saying but the point is that there are cases in which manual alignment of the display black point to input code 0 achieves better results than what dispcal is producing. So how do you handle that situation? The user can reset brightness after dispcal to remove any measurable above black offset or hack the .cal file, but neither of these options seems satisfactory. The algorithm could be improved but's in unclear where the problem lies, maybe the algorithm is fine and quantization in the video chain or measurements limits the performance.
I agree. What I was trying to get at is that adjusting a display visually by tweaking the controls and trying to make input code 0 distinct from input code 1 may be tricky to do depending on the display, its settings and the situation. If anything, this is actually in favor of the way dispcal works currently (i.e. don't make any assumptions about the display, measure and create the curves accordingly. Which for my particular TV works very well I might add, even after calibration input code 0 stays at 0, there's no clipping, and the progression out of black is smooth).

Last edited by fhoech; 09-11-2014 at 11:12 AM.
fhoech is offline  
post #2619 of 2785 Old 09-11-2014, 11:19 AM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 901
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 92 Post(s)
Liked: 75
Quote:
Originally Posted by gwgill View Post
But the 256 entries are not measured or created explicitly. Far fewer points are measured, and then the 256 entries are generated by interpolation.
understood, thanks.
spacediver is online now  
post #2620 of 2785 Old 09-12-2014, 01:04 AM
Senior Member
 
James Freeman's Avatar
 
Join Date: Sep 2013
Posts: 216
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 128 Post(s)
Liked: 57
The i1D3 seems to drift in temperature of about 400K in a duration of 30min when on a warm screen (CCFL or plasma) despite it being internally compensated.

I have asked this in the past, whether I should wait till the i1D3 warms up on screen and only then calibrate or the meter should be at room temp' and calibrate from a tripod.
ConnecTEDDD answered that the noise floor of the i1D3 sensor raises with temp, and it should be always at room temp to be accurate especially on hot displays like plasma.

Quote:
Originally Posted by ConnecTEDDD
when the meter is becoming warm, more noise is added to the sensor. (reducing S/N ratio)
To mend this while still using contact calibration,
I use "interactive display adjustment" to adjust the white point to 6500K with the meter at room temp' and 10cm away from the screen.
Then I cancel the current calibration because there is no option to set the white point without continuing to calibration,
and start a new one with the meter on screen but with both black and white drift compensation.
This gives me nice and even temp' across the greyscale thanks to argyllCMS's drift compensation.

I want to make sure everybody is aware of this.
James Freeman is offline  
post #2621 of 2785 Old 09-12-2014, 06:29 AM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
Quote:
Originally Posted by James Freeman View Post
The i1D3 seems to drift in temperature of about 400K in a duration of 30min when on a warm screen (CCFL or plasma) despite it being internally compensated.
I've not personally experienced this on my own displays (white chromaticity measurements are very repeatable on my WLED TV, and also on my CCFL wide-gamut monitor after warm-up of around 30-40 minutes for the latter), although I could imagine that some displays may get considerably warmer than what I have.

Quote:
Originally Posted by James Freeman View Post
I use "interactive display adjustment" to adjust the white point to 6500K with the meter at room temp' and 10cm away from the screen.
Then I cancel the current calibration because there is no option to set the white point without continuing to calibration,
and start a new one with the meter on screen but with both black and white drift compensation.
Wouldn't enabling drift compensation right at the start and then simply putting the instrument on the screen after adjusting the whitepoint also work (I think dispcal ignores drift compensation during adjustment)?
fhoech is offline  
post #2622 of 2785 Old 09-12-2014, 07:09 AM
Senior Member
 
James Freeman's Avatar
 
Join Date: Sep 2013
Posts: 216
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 128 Post(s)
Liked: 57
Quote:
Originally Posted by fhoech View Post
I've not personally experienced this on my own displays (white chromaticity measurements are very repeatable on my WLED TV, and also on my CCFL wide-gamut monitor after warm-up of around 30-40 minutes for the latter), although I could imagine that some displays may get considerably warmer than what I have.
My screen is Dell U2410 (Wide Gamut CCFL) with brightness at 0 giving 80 cd/m2 which is plenty bright for everything I do.
The temp of the screen is not that warm, its a little warmer than room temp (by hand) but still the meter drifts.
Judging by that behavior I can tell that the i1D3 is very temp sensitive and I should take that into account on my next calibrations especially my plasma.

To test that:
The i1D3 warms up for about 30 min on the table in room temp.
Put the room temp meter on screen (screen already warmed up) and immediately measure white point, wait 1-2 hours with meter still on screen and measure the white point again.
Mine drifts colder about 400-500K.

Quote:
Originally Posted by fhoech View Post
Wouldn't enabling drift compensation right at the start and then simply putting the instrument on the screen after adjusting the whitepoint also work (I think dispcal ignores drift compensation during adjustment)?
If it's true then Yes.

Last edited by James Freeman; 09-12-2014 at 07:30 AM.
James Freeman is offline  
post #2623 of 2785 Old 09-12-2014, 08:26 AM
Advanced Member
 
-Hitman-'s Avatar
 
Join Date: Nov 2004
Location: Manchester, UK
Posts: 678
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 42 Post(s)
Liked: 63
Quote:
Originally Posted by James Freeman View Post
The i1D3 seems to drift in temperature of about 400K in a duration of 30min when on a warm screen (CCFL or plasma) despite it being internally compensated.

I have asked this in the past, whether I should wait till the i1D3 warms up on screen and only then calibrate or the meter should be at room temp' and calibrate from a tripod.

Most folk use the i1d3 "off " screen to combat the problems with heat, also the readings will suffer from noise using contact mode with plasma's.
-Hitman- is offline  
post #2624 of 2785 Old 09-12-2014, 09:09 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,725
Mentioned: 30 Post(s)
Tagged: 0 Thread(s)
Quoted: 258 Post(s)
Liked: 361
Quote:
Originally Posted by fhoech View Post
I've not personally experienced this on my own displays (white chromaticity measurements are very repeatable on my WLED TV, and also on my CCFL wide-gamut monitor after warm-up of around 30-40 minutes for the latter), although I could imagine that some displays may get considerably warmer than what I have.
I've not had stability (or noise) problems on my plasma in contact mode either. I did test this awhile back after letting the display warm up for an hour (probe was stowed away from display and disconnected). Then started monitoring the output over the course of an hour. The probe reported CCT drifted about 20 kelvin over the first 15 minutes and then was rock solid.

zoyd is online now  
post #2625 of 2785 Old 09-12-2014, 10:38 AM
Advanced Member
 
-Hitman-'s Avatar
 
Join Date: Nov 2004
Location: Manchester, UK
Posts: 678
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 42 Post(s)
Liked: 63
Interesting, thanks zoyd.


Personally I've had problems with erratic readings from an i1pro on my krp-600m, which generates a lot of heat on it's surface especially being 60", so used non contact mode since with no problems thereafter.


However with the i1d3 I've used non contact from the start, especially as it was recommended to do so by a few users I know.


Also I've been told prolonged contact/heat can aid/speed up drifting within the sensors filters.


Did you only test with 1 display (VT?) and the D3 or other display models/sensors?

Last edited by -Hitman-; 09-12-2014 at 10:49 AM. Reason: added info
-Hitman- is offline  
post #2626 of 2785 Old 09-12-2014, 11:43 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,725
Mentioned: 30 Post(s)
Tagged: 0 Thread(s)
Quoted: 258 Post(s)
Liked: 361
Quote:
Originally Posted by -Hitman- View Post
Interesting, thanks zoyd.


Personally I've had problems with erratic readings from an i1pro on my krp-600m, which generates a lot of heat on it's surface especially being 60", so used non contact mode since with no problems thereafter.
yes, the i1pro is more sensitive to temperature changes.

Quote:
However with the i1d3 I've used non contact from the start, especially as it was recommended to do so by a few users I know.
Nothing wrong with being cautious, I just like to point out cases where potential problems are exaggerated, especially when lacking any measurements.

Quote:
Also I've been told prolonged contact/heat can aid/speed up drifting within the sensors filters.
I've also not seen any problem with long-term stability over the past year with very heavy use, so I would not worry about occasional routine calibrations in contact mode.

Quote:
Did you only test with 1 display (VT?) and the D3 or other display models/sensors?
That was a 51" Samsung but it generates plenty of heat. The only other sensors I use routinely are the i1pro2 and JETI which are always tripod mounted.
-Hitman- likes this.
zoyd is online now  
post #2627 of 2785 Old 09-12-2014, 04:34 PM
Member
 
brucek's Avatar
 
Join Date: Feb 2002
Location: SF Bay Area
Posts: 35
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
The current DispcalGUI 2.5 is no longer an exact match for the detailed instructions at the start of this thread. As a novice facing a daunting learning curve already, I'd hate to be messing myself up right from the start by guessing wrong at what the new inputs should be. Would someone mind giving the updated instructions as they apply to 2.5?

The differences are:

1. Calibration setting: Tone Curve: no longer offers the ability to specify gamma 2.4 / absolute. Does this require a compensating adjustment anywhere else?

2. In profile settings, there is no longer an option to disable Smooth B2A tables. Does this require a compensating adjustment anywhere else?

3. In profile settings, there is no longer an option to enable Low quality B2A tables. Does this require a compensating adjustment anywhere else?

4. In edit testchart, there are now three gamma-specific Rec709 preconditioning profiles vs one generic. Which gamma should be selected? Does the choice have to be matched in any other gamma specific inputs?

5. In Create 3D LUT, there are now three gamma-specific Rec709 source profiles. Which gamma should be selected? Does the choice have to be matched in any other gamma specific inputs?

Thanks!
brucek is offline  
post #2628 of 2785 Old 09-13-2014, 01:23 AM
Advanced Member
 
mightyhuhn's Avatar
 
Join Date: Nov 2013
Posts: 648
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 269 Post(s)
Liked: 94
Quote:
Originally Posted by brucek View Post
The current DispcalGUI 2.5 is no longer an exact match for the detailed instructions at the start of this thread. As a novice facing a daunting learning curve already, I'd hate to be messing myself up right from the start by guessing wrong at what the new inputs should be. Would someone mind giving the updated instructions as they apply to 2.5?

The differences are:

1. Calibration setting: Tone Curve: no longer offers the ability to specify gamma 2.4 / absolute. Does this require a compensating adjustment anywhere else?

2. In profile settings, there is no longer an option to disable Smooth B2A tables. Does this require a compensating adjustment anywhere else?

3. In profile settings, there is no longer an option to enable Low quality B2A tables. Does this require a compensating adjustment anywhere else?

4. In edit testchart, there are now three gamma-specific Rec709 preconditioning profiles vs one generic. Which gamma should be selected? Does the choice have to be matched in any other gamma specific inputs?

5. In Create 3D LUT, there are now three gamma-specific Rec709 source profiles. Which gamma should be selected? Does the choice have to be matched in any other gamma specific inputs?

Thanks!
1: choice bt 1886 like in the guide here and it does. nothing changed at all just more extra options leave them at default.

2: if you choice madVrRas preset it is active by default it is under advanced options right next to profile type. the default settings are already set right.
3: see 2

4: the guide shows genatic so don't choice one with a gamma. as far as I know they were always there.

5: see 4


by the way the default madVR preset settings are totally fine. they were created to create a madVR 3D LUT.
brucek likes this.
mightyhuhn is offline  
post #2629 of 2785 Old 09-13-2014, 06:18 AM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
The instructions in the first post regarding dispcalGUI settings could probably be condensed from 16 to 9 steps (including optional steps):

  1. Select "madVR" under "Settings".
  2. (Optional) If you want to use iterative gray balance calibration (may improve gray balance, but may also slightly raise your black level), set calibration tone curve from "As measured" to "Rec. 1886".
  3. (Optional) If you want to use a larger testchart for potential higher accuracy, select any of the testcharts for LUT profiles optimized for gamma 2.2 with Rec. 709 primaries (default size is XXL with 1399 patches), or use the testchart editor to create your own as outlined in the first post.
  4. Click "Calibrate & profile".
  5. After the process is complete, a window will show the profile self check error as well as gamut coverage. The self check error should not be higher than 0.5 DE average and 5 DE peak. Click "Create 3D LUT...".
  6. Ensure that Rec709.icm is selected as source profile and "Apply tone response curve" is enabled (those are the defaults, so unless changed deliberately those will already be set correctly).
  7. (Optional) Adjust tone response curve options as desired. Default is Rec. 1886 (input offset). To target a different curve, e.g. what people usually refer to as "pure power" gamma curve, set gamma from "Absolute" to "Relative" and 100% output offset. A blend between input offset and output offset can be achieved by setting a value between 0% and 100%.
  8. Select "madVR (.3dlut)" as 3D LUT file format.
  9. Click "Create 3D LUT...".
brucek likes this.
fhoech is offline  
post #2630 of 2785 Old 09-13-2014, 06:49 PM
Member
 
baii's Avatar
 
Join Date: Jul 2013
Posts: 18
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
I just tried to use the madvr preset in dispcalgui 2.5 with 1.7 argyll beta, it does not initiate the dispcal command but go to colorprof directly instead, is it intended?
baii is offline  
post #2631 of 2785 Old 09-13-2014, 06:52 PM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
Quote:
Originally Posted by baii View Post
I just tried to use the madvr preset in dispcalgui 2.5 with 1.7 argyll beta, it does not initiate the dispcal command but go to colorprof directly instead, is it intended?
Yes. See my above post.
fhoech is offline  
post #2632 of 2785 Old 09-13-2014, 09:33 PM
Member
 
brucek's Avatar
 
Join Date: Feb 2002
Location: SF Bay Area
Posts: 35
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
Thanks guys for the clarifications. Its helpful to know what you don't have to worry about ;-)

I know you're both telling me that there is a generic Rec709 profile, but I haven't been able to find it, even after using tools -> restore defaults. When I look at, for example, Create 3D LUT -> Source Profile, the Rec 709 choices I see are:

Rec709_Gamma22.icm
Rec709_Gamma235.icm
Rec709_Gamma24.icm

There's other choices for Rec601 and SMTPE240M and ACES, but nothing else that includes Rec709.

I got fairly pleasing results first time out using the iterative gray balance calibration + basic 2000 test charts + gamma 2.2 versions of Rec709, but I am noticing that raised black level. I'll try again without the iterative gray balance and see which I prefer. Thanks for the tips!
brucek is offline  
post #2633 of 2785 Old 09-14-2014, 02:04 AM
Advanced Member
 
mightyhuhn's Avatar
 
Join Date: Nov 2013
Posts: 648
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 269 Post(s)
Liked: 94
you can find the rec 709 in argyllCMS folder under ref.
brucek likes this.
mightyhuhn is offline  
post #2634 of 2785 Old 09-14-2014, 05:36 AM
Member
 
Francois76l's Avatar
 
Join Date: May 2014
Location: FRANCE
Posts: 104
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 29 Post(s)
Liked: 10
Kuro FPJ-9000 3DLUT calibration

Yesterday night, I calibrated a Kuro FPJ9000 with dispcalgui. Me and the owner of the projector were totally amazed by the results!!!
DispcalGUI and Argyll CMS are really fantastics tools! Huge thanks for the developpers!

Sorry If I put colorhcfr screenshots but I find it more readable. Especially the CIE diagram
Attached Thumbnails
Click image for larger version

Name:	Echelle_Gris.png
Views:	47
Size:	54.0 KB
ID:	261834   Click image for larger version

Name:	CIE.png
Views:	61
Size:	379.4 KB
ID:	261842   Click image for larger version

Name:	Gamma.png
Views:	51
Size:	42.9 KB
ID:	261850   Click image for larger version

Name:	Luminance.png
Views:	49
Size:	48.4 KB
ID:	261858   Click image for larger version

Name:	Niveau_RVB.png
Views:	49
Size:	51.1 KB
ID:	261866  

Click image for larger version

Name:	Primaires.png
Views:	49
Size:	89.2 KB
ID:	261874   Click image for larger version

Name:	Temp_Colo.png
Views:	45
Size:	39.0 KB
ID:	261882  

HiFi:Nad C320BEE+Marantz CD-63 MkII-Ki Signature+Boston VR2+DIY cabling
HC:Onkyo TX-SR875+Boston VR2+CR8+Chorus cc800v+Subwoofer MJ Acoustic Ref1 MKII
DISPLAYS:Panasonic P50ST50E/JVC DLA-X35+110" gain 0.8+HTPC (LAV+ffdshow+Avisynth+MPC-BE+MadVr)
Francois76l is offline  
post #2635 of 2785 Old 09-14-2014, 05:42 AM
Member
 
Francois76l's Avatar
 
Join Date: May 2014
Location: FRANCE
Posts: 104
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 29 Post(s)
Liked: 10
I used the correction matrice given by Icaro and it was just perfect (UHP lamp)
Quote:
Originally Posted by Icaro View Post

I do not know what JVC model you have,
but also in this case (for what it's worth),
this is my matrix correction for the i1D3 made with a i1Pro (rev.D), and my JVC projector (HD990),
in an environment dedicated and totally obscured.
is present within the .THC for HCFR and the .CCMX for DispcalGui,
always use them both (if applicable to your vpj)

i1D3-MatrixCorrectionforJVCHD990i1Pro.zip 1k .zip file

HiFi:Nad C320BEE+Marantz CD-63 MkII-Ki Signature+Boston VR2+DIY cabling
HC:Onkyo TX-SR875+Boston VR2+CR8+Chorus cc800v+Subwoofer MJ Acoustic Ref1 MKII
DISPLAYS:Panasonic P50ST50E/JVC DLA-X35+110" gain 0.8+HTPC (LAV+ffdshow+Avisynth+MPC-BE+MadVr)
Francois76l is offline  
post #2636 of 2785 Old 09-14-2014, 08:23 AM
Newbie
 
dansrfe's Avatar
 
Join Date: Nov 2008
Posts: 3
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 0
Quote:
Originally Posted by fhoech View Post
The instructions in the first post regarding dispcalGUI settings could probably be condensed from 16 to 9 steps (including optional steps):

  1. Select "madVR" under "Settings".
  2. (Optional) If you want to use iterative gray balance calibration (may improve gray balance, but may also slightly raise your black level), set calibration tone curve from "As measured" to "Rec. 1886".
  3. (Optional) If you want to use a larger testchart for potential higher accuracy, select any of the testcharts for LUT profiles optimized for gamma 2.2 with Rec. 709 primaries (default size is XXL with 1399 patches), or use the testchart editor to create your own as outlined in the first post.
  4. Click "Calibrate & profile".
  5. After the process is complete, a window will show the profile self check error as well as gamut coverage. The self check error should not be higher than 0.5 DE average and 5 DE peak. Click "Create 3D LUT...".
  6. Ensure that Rec709.icm is selected as source profile and "Apply tone response curve" is enabled (those are the defaults, so unless changed deliberately those will already be set correctly).
  7. (Optional) Adjust tone response curve options as desired. Default is Rec. 1886 (input offset). To target a different curve, e.g. what people usually refer to as "pure power" gamma curve, set gamma from "Absolute" to "Relative" and 100% output offset. A blend between input offset and output offset can be achieved by setting a value between 0% and 100%.
  8. Select "madVR (.3dlut)" as 3D LUT file format.
  9. Click "Create 3D LUT...".
If I plan to use the resulting profile via dispwin, should "Low quality PCS-to-device tables" be unchecked?

And if so, should "Enhance effective resolution of colormetric PCS-to-device table" be checked and set to something higher than "33x33x33"? What about checking "Smoothing" and changing the "Default rendering intent to "Absolute colormetric"?

Also, Calibration & Profile must be clicked for calibration curves to be embedded in the profile, correct? Thanks!

Last edited by dansrfe; 09-14-2014 at 08:42 AM.
dansrfe is offline  
post #2637 of 2785 Old 09-14-2014, 01:15 PM
Senior Member
 
fhoech's Avatar
 
Join Date: Apr 2013
Location: Stuttgart, Germany
Posts: 222
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 53 Post(s)
Liked: 44
Quote:
Originally Posted by dansrfe View Post
If I plan to use the resulting profile via dispwin, should "Low quality PCS-to-device tables" be unchecked?

And if so, should "Enhance effective resolution of colormetric PCS-to-device table" be checked
In that case, yes to both.

Quote:
Originally Posted by dansrfe View Post
and set to something higher than "33x33x33"?
Not necessarily. The default matches the device-to-PCS table resolution.

Quote:
Originally Posted by dansrfe View Post
What about checking "Smoothing"
It shouldn't harm to have it enabled.

Quote:
Originally Posted by dansrfe View Post
and changing the "Default rendering intent to "Absolute colormetric"?
This won't make any difference.

Quote:
Originally Posted by dansrfe View Post
Also, Calibration & Profile must be clicked for calibration curves to be embedded in the profile, correct?
You can also use an existing calibration (if you have one), in which case you'd use "Profile only".
fhoech is offline  
post #2638 of 2785 Old 09-16-2014, 04:21 AM
Newbie
 
bsod's Avatar
 
Join Date: Jun 2006
Posts: 9
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 0
Hello Everyone!

My previous attempts for using madvr with 3dlut (using an old guide) introduced some increased black level biased to blue and some posterisation. Overall it was not a pleasant picture. Decided to ditch it then.
This guide seems more refined and my new projector is about to arrive, I'm going to give one more try for it.

I have a Peroni focused silver screen that makes harder to read consistent data with my i1pro spectrophotometer.
If I'm correct the device is not good at darker regions. Is a way to increase sampling in these regions, to avoid posterisation? Using smoothing helps, or I'm better off getting a colorimeter?

Thanks!
bsod is online now  
post #2639 of 2785 Old 09-16-2014, 09:14 AM
Advanced Member
 
-Hitman-'s Avatar
 
Join Date: Nov 2004
Location: Manchester, UK
Posts: 678
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 42 Post(s)
Liked: 63
Quote:
Originally Posted by bsod View Post
Hello Everyone!

My previous attempts for using madvr with 3dlut (using an old guide) introduced some increased black level biased to blue and some posterisation. Overall it was not a pleasant picture. Decided to ditch it then.
This guide seems more refined and my new projector is about to arrive, I'm going to give one more try for it.

I have a Peroni focused silver screen that makes harder to read consistent data with my i1pro spectrophotometer.
If I'm correct the device is not good at darker regions. Is a way to increase sampling in these regions, to avoid posterisation? Using smoothing helps, or I'm better off getting a colorimeter?

Thanks!

You have a couple of options to deal with the lower grey area, dark region emphasis, black point compensation, increasing the number of white patches in the testchart and output offsets


I would recommend getting an i1d3 and combining it/profiling with your i1pro to create an offset for the D3, then you can get much better results below 30% grey than you can with the i1pro alone.
-Hitman- is offline  
post #2640 of 2785 Old 09-17-2014, 04:31 AM
Member
 
LucaT's Avatar
 
Join Date: Mar 2014
Posts: 28
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 10
Hello guys, I used DispcalGUI 2.5 and Argyll 1.6.3 and followed all steps for 3dlut creation, but:

- if I do "only calibration" I create the 3dlut, enable it in MadVR and when I check results with HCFR gamma and grey-scale results completely unbalanced

- if I do "calibration & profile" it stops after more than one hour showing error
"Dispread.exe could not be started"

Last edited by LucaT; 09-17-2014 at 10:04 AM.
LucaT is offline  
Reply Display Calibration

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

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
Trackbacks are Off
Pingbacks are Off
Refbacks are Off