HCFR - Open source projector and display calibration software - Page 154 - AVS | Home Theater Discussions And Reviews
Forum Jump: 
 
Thread Tools
Old 10-13-2014, 12:19 PM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,743
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 162
Quote:
Originally Posted by pisymbol View Post
What's strange is most of CMS settings were very slight OTHER THAN Cyan. If it wasn't for Cyan, I'd say the whole calibration was fairly smooth other than dealing with HCFR crashing a lot during the CMS part.
Specifically, which of the adjustment(s) are large?

Quote:

Note, the Vizio P-Series CMS controls seem to be useless at 100% saturation (confirmed by two other more experienced calibrators).
This is fairly normal for all flat-panels, and CRTs ... it's very hard to change the chromaticity of the 100% physical primaries ... unless you have a wide native gamut.

PS: One would expect a 4k/UHD display to have a significantly wider gamut than REC709 ... so apparently either Vizio doesn't trust us with the CMS, or ...

Oh well, I guess it doesn't really matter since there's no consumer REC2020(ish) content yet.

Last edited by HDTVChallenged; 10-13-2014 at 12:35 PM.
HDTVChallenged is offline  
Sponsored Links
Advertisement
 
Old 10-13-2014, 01:03 PM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by HDTVChallenged View Post
Specifically, which of the adjustment(s) are large?
Cyan saturation. I had to bump it to the maximum to achieve a 1.3 dE. The rest of the colors were relatively minor adjustments and all under 1.0 dE.
pisymbol is offline  
Old 10-13-2014, 01:20 PM
Senior Member
 
Andrei_VVB's Avatar
 
Join Date: Mar 2010
Posts: 200
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 17
Maybe because you've targeted such a high max luminance you have clipped cyan?
Is that possible?
Andrei_VVB is offline  
Old 10-13-2014, 03:03 PM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by Andrei_VVB View Post
Maybe because you've targeted such a high max luminance you have clipped cyan?
Is that possible?
Hmmm... are you saying if I targeted 30-40 ftL, I would have had less of an issue with CMS?

I had thought for a daytime setting, 55ftL would not be so bad. But perhaps that is even too bright.

Last edited by pisymbol; 10-13-2014 at 03:16 PM.
pisymbol is offline  
Old 10-13-2014, 03:22 PM
Senior Member
 
Andrei_VVB's Avatar
 
Join Date: Mar 2010
Posts: 200
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 17
It is just a question. Sometimes rising contrast too much might induce some clipping.
Some more knowledgeable folks around here might say if that's possible.
Do that night calibration and see how it goes.
Andrei_VVB is offline  
Old 10-13-2014, 09:48 PM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,743
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 162
Quote:
Originally Posted by pisymbol View Post
Cyan saturation. I had to bump it to the maximum to achieve a 1.3 dE. The rest of the colors were relatively minor adjustments and all under 1.0 dE.
Humm ... interesting. One might also expect to find an issue with green (or blue) in this case. It's definitely odd, but so long as hue and brightness look ok, I wouldn't stress over it. Your dE's look fine. As Andrei suggested, the next step would be to run the color-checker measures.
HDTVChallenged is offline  
Old 10-13-2014, 10:51 PM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by HDTVChallenged View Post
Humm ... interesting. One might also expect to find an issue with green (or blue) in this case. It's definitely odd, but so long as hue and brightness look ok, I wouldn't stress over it. Your dE's look fine. As Andrei suggested, the next step would be to run the color-checker measures.
I just finished my Night calibration at 35ish ftL (I got to tell you, this TV is bright!).

Well, the fact is I can play with tint to bring Cyan down to a bit under a dE 1.0 but it's really not worth it (then blue and green need massive adjustments). I decided against it and left it at zero.

The rest are definitely inline with what I expect. I need to run GCD and color-checker just to see how off it is, but tonight night is dialed in.

AFAICT, this Vizio P502uiB1 just has trouble with Cyan for whatever reason. I am hoping a professional calibrator with a real honest to goodness spectraradiometer can confirm it.
pisymbol is offline  
Old 10-13-2014, 11:41 PM
Advanced Member
 
orion2001's Avatar
 
Join Date: Jun 2008
Posts: 728
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 448 Post(s)
Liked: 299
Quote:
Originally Posted by pisymbol View Post
I just finished my Night calibration at 35ish ftL (I got to tell you, this TV is bright!).

Well, the fact is I can play with tint to bring Cyan down to a bit under a dE 1.0 but it's really not worth it (then blue and green need massive adjustments). I decided against it and left it at zero.

The rest are definitely inline with what I expect. I need to run GCD and color-checker just to see how off it is, but tonight night is dialed in.

AFAICT, this Vizio P502uiB1 just has trouble with Cyan for whatever reason. I am hoping a professional calibrator with a real honest to goodness spectraradiometer can confirm it.
You need to work on your Gamma quite a bit. It looks like you are matching RGB to get neutral whitebalance, but you are not paying attention to your luminance targets in HCFR in order to hit the BT.1886 target Gamma curve. If you have missed it so far, there is actually a row in the HCFR grayscale table for Target Luminance that is typically hidden but can be scrolled down to. You need to use the 10pt whitebalance controls to start dialing the measured Y value to the target Y value for each point in order to try and better match your Gamma curve.

Also, I'd suggest doing some saturation sweep measurements before you spend too much time dialing in the primary and secondary colors at a specific saturation levels. The saturation sweeps should give you a more holistic view of the panel color performance and should help you decide on what strategy to use to achieve a good overall color calibration.

-------------------------------------------------------------------------
Recommended calibration settings for Samsung PN60F5300B
orion2001 is offline  
Old 10-14-2014, 06:13 AM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by orion2001 View Post
You need to work on your Gamma quite a bit. It looks like you are matching RGB to get neutral whitebalance, but you are not paying attention to your luminance targets in HCFR in order to hit the BT.1886 target Gamma curve. If you have missed it so far, there is actually a row in the HCFR grayscale table for Target Luminance that is typically hidden but can be scrolled down to. You need to use the 10pt whitebalance controls to start dialing the measured Y value to the target Y value for each point in order to try and better match your Gamma curve.

Also, I'd suggest doing some saturation sweep measurements before you spend too much time dialing in the primary and secondary colors at a specific saturation levels. The saturation sweeps should give you a more holistic view of the panel color performance and should help you decide on what strategy to use to achieve a good overall color calibration.
I thought I match the bars and luminance is taking care for me, no? My gamma is 2.19, target was 2.20. I'm confused orion2001: I thought that was pretty good?

EDIT: I see now HCL on the bottom, but how do I match luminance outside of RGB using HCFR?
pisymbol is offline  
Old 10-14-2014, 06:39 AM
Advanced Member
 
orion2001's Avatar
 
Join Date: Jun 2008
Posts: 728
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 448 Post(s)
Liked: 299
Quote:
Originally Posted by pisymbol View Post
I thought I match the bars and luminance is taking care for me, no? My gamma is 2.19, target was 2.20. I'm confused orion2001: I thought that was pretty good?

EDIT: I see now HCL on the bottom, but how do I match luminance outside of RGB using HCFR?
In the Gamma curve you posted, the Yellow curve is your measured Gamma. The white curve is your target BT.1886 Gamma curve. If you want to target a Flat power law Gamma curve instead of BT.1886 (I prefer BT.1886), then you need to first change the target Gamma curve in the Options.

In order to meet your luminance targets, you will need to tweak your 10pt white balance controls (and 2 pt controls if you have them). Firstly, remember that you should not tweak your 100% and 0% settings, when you are trying to match your gamma curve as the gamma curve depends on the 100% and 0% readings. So, you first set the 100% reading to a neutral white and use contrast to set it to an acceptable brightness target (~35fL for dark room viewing, ~45-55fL for bright room).

If you look at your Luminance curve, you will see that pretty much across the 50-90% range you are higher than the target luminance. If you have 2 pt control, I would first use the 2pt control at the high end to dial back R, G and B in equal amounts to maintain a neutral white but to lower your luminance overall in the 80% region to get close to the luminance target for 80%. Then start working with 90% 10pt control. Lower RGB in equal proportions to lower overall luminance to match the target luminance while maintaining a neutral white. Repeat across the board for all the points that you can tweak with your 10pt white balance controls. Then redo your grayscale measurement and you should be much closer to target Gamma curve. You may need another pass with very minor tweaks to really refine your gamma curve to match the target curve.

At the moment, your dE formula that is being used (in the Advanced or Options menu) probably only values chromaticity and does not factor in how far you are from your luminance targets, which is why your dE looks excellent (due to all measurements being close to neutral white) even though you are quite off on your luminance targets. I believe there are other formulas that can be selected in HCFR options that account for luminance in calculation of reported dE values.

-------------------------------------------------------------------------
Recommended calibration settings for Samsung PN60F5300B
orion2001 is offline  
Old 10-14-2014, 08:02 AM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by orion2001 View Post
In the Gamma curve you posted, the Yellow curve is your measured Gamma. The white curve is your target BT.1886 Gamma curve. If you want to target a Flat power law Gamma curve instead of BT.1886 (I prefer BT.1886), then you need to first change the target Gamma curve in the Options.
No, BT.1886 is what I'd like to achieve.

EDIT: I confirmed BT.1886 is set as the reference during my calibration.

Quote:
In order to meet your luminance targets, you will need to tweak your 10pt white balance controls (and 2 pt controls if you have them). Firstly, remember that you should not tweak your 100% and 0% settings, when you are trying to match your gamma curve as the gamma curve depends on the 100% and 0% readings. So, you first set the 100% reading to a neutral white and use contrast to set it to an acceptable brightness target (~35fL for dark room viewing, ~45-55fL for bright room).

If you look at your Luminance curve, you will see that pretty much across the 50-90% range you are higher than the target luminance. If you have 2 pt control, I would first use the 2pt control at the high end to dial back R, G and B in equal amounts to maintain a neutral white but to lower your luminance overall in the 80% region to get close to the luminance target for 80%. Then start working with 90% 10pt control. Lower RGB in equal proportions to lower overall luminance to match the target luminance while maintaining a neutral white. Repeat across the board for all the points that you can tweak with your 10pt white balance controls. Then redo your grayscale measurement and you should be much closer to target Gamma curve. You may need another pass with very minor tweaks to really refine your gamma curve to match the target curve.

At the moment, your dE formula that is being used (in the Advanced or Options menu) probably only values chromaticity and does not factor in how far you are from your luminance targets, which is why your dE looks excellent (due to all measurements being close to neutral white) even though you are quite off on your luminance targets. I believe there are other formulas that can be selected in HCFR options that account for luminance in calculation of reported dE values.
Here is what I did:

1) Insert AVS709 disc -> Basic Settings
2) Set Brightness and Contrast appropriately using the flashing bars
3) **** Set Backlight to meet a luminous target of 35ftL using the 100% white pattern (wait...maybe I should use the 100% gray instead? Is that where I screwed up? Since gamma is a function from black->white, I thought 100% white would be OK)
4) Adjust 80% and 30% 2pt controls to arrive at the lowest possible dE
5) Adjust 11pt controls (5%-10%-20%...100%) to arrive at the lowest possible dE
6) Go back to basic settings and verify that the right bars are still blinking, if not, rinse, repeat

What am I doing wrong and also how can I see how off I am in luminous while measuring? I assume I want my yellow line to be flatter against the BT.1886 white reference one so my average is inline with the reference.

orion, I really appreciate the feedback. I want to get this right.

NOTE: If you look at the Luminance curve, it looks pretty good to me though.

ONE thing I noticed is this set as a gamma setting. It's currently at 2.2, perhaps it should be at 2.4 for BT.1886?

Last edited by pisymbol; 10-14-2014 at 08:27 AM.
pisymbol is offline  
Old 10-14-2014, 09:04 AM
Senior Member
 
carillon's Avatar
 
Join Date: Nov 2006
Location: Knoxville, TN
Posts: 243
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 34 Post(s)
Liked: 10
I have really enjoyed this conversation including pisymbol and others who are pitching in to help. I'm a rookie to HCFR but am learning a great deal. I would desperately like to have access to the "guide" for using HCFR to properly calibrate an LED TV. Where can I find this? Thanks!
carillon is offline  
Old 10-14-2014, 09:07 AM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by carillon View Post
I have really enjoyed this conversation including pisymbol and others who are pitching in to help. I'm a rookie to HCFR but am learning a great deal. I would desperately like to have access to the "guide" for using HCFR to properly calibrate an LED TV. Where can I find this? Thanks!
Here.

Though if I get enough help, I think we should start creating a Wiki page. Wondering what the HCFR group things of that? I am willing to document my entire calibration and all the issues I went through (as well update the guide against AVS709 and GCD).

But first, I want to get my gamma right....help!
pisymbol is offline  
Old 10-14-2014, 09:45 AM
Advanced Member
 
orion2001's Avatar
 
Join Date: Jun 2008
Posts: 728
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 448 Post(s)
Liked: 299
Quote:
Originally Posted by pisymbol View Post
Here is what I did:

1) Insert AVS709 disc -> Basic Settings
2) Set Brightness and Contrast appropriately using the flashing bars
3) **** Set Backlight to meet a luminous target of 35ftL using the 100% white pattern (wait...maybe I should use the 100% gray instead? Is that where I screwed up? Since gamma is a function from black->white, I thought 100% white would be OK)
4) Adjust 80% and 30% 2pt controls to arrive at the lowest possible dE
5) Adjust 11pt controls (5%-10%-20%...100%) to arrive at the lowest possible dE
6) Go back to basic settings and verify that the right bars are still blinking, if not, rinse, repeat

What am I doing wrong and also how can I see how off I am in luminous while measuring? I assume I want my yellow line to be flatter against the BT.1886 white reference one so my average is inline with the reference.
As such what you are doing is pretty much spot on. The problem is that by default, HCFR calculates dE for grayscale without taking into account Gamma curve target. As a result, when you are making your 2 and 10 pt white balance adjustments, all you are doing are ensuring that the white chromaticity (xy coordinate) is accurate to D65 target (i.e., neutral white). However, minimizing dE does not ensure that you are also matching your luminance target as desired by your reference Gamma curve of choice.

In order to also match your target gamma curve correctly, there are a couple of options:

1) Go to preferences->Advanced and set Grayscale dE handling to "Absolute Y w/ Gamma". I haven't used this myself, but this should in theory now take your luminance target into account for the dE calculations. You will find large errors if you now do a sweep and you will need to lower RGB values till you are at D65 AND also meet the required luminance target for your current gray level as determined by your Gamma curve choice in order to minimize dE.

2) The other option is to keep doing as you are doing. Now that you have done a grayscale measurement, you can look at the grayscale table values for each step from 10 - 90%. Then focus on each level and dial in the target luminance. For example: Let's look at your 90% Y value. It was measured to be 113 cd/m2. If you scroll down in the table view (or you can click the "down arrow" in the little widget next to the "Editable Data" checkbox to reveal lower lines in the table) you will see the Target Y also in cd/m2. Let's say this reads 108 cd/m2. This means that your display is outputting too much light at 90 IRE. So now highlight the 90 IRE column, then select free measure and display the 90 IRE pattern on the display. Keep tweaking RGB down till you still have all bars showing 100% (i.e. you have a neutral white) but your measured Y is now lowered to 108 (or close to it). Repeat the process for each IRE step and you will end up with neutral whites and your luminance curve will match your target curve as determined by your reference gamma curve selection.

If everything pans out correctly, your gamma curve should look like the gamma curve in my signature (Click the "Spoiler" tag in that post to reveal my full calibration with all the calibration curves)

Quote:
orion, I really appreciate the feedback. I want to get this right.

NOTE: If you look at the Luminance curve, it looks pretty good to me though.
Unfortunately, the luminance curve can be misleading as even big errors in gamma can look very small in that plot. The main curve of interest is the gamma curve which shows how far off the luminance target you are for each step increment in white/gray. As you can see in your curve, your measured gamma is substantially off the target gamma. The average gamma should not be something you should bother about. A lot of poor calibrations by folks are due to them assuming that matching average gamma is good enough. With a 10pt control, there is no reason you shouldn't be able to pretty much nail the target gamma curve.

Quote:
ONE thing I noticed is this set as a gamma setting. It's currently at 2.2, perhaps it should be at 2.4 for BT.1886?
I'd leave it as is for now.

Good luck!

-------------------------------------------------------------------------
Recommended calibration settings for Samsung PN60F5300B
orion2001 is offline  
Old 10-14-2014, 10:45 AM
Senior Member
 
carillon's Avatar
 
Join Date: Nov 2006
Location: Knoxville, TN
Posts: 243
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 34 Post(s)
Liked: 10
Quote:
Originally Posted by pisymbol View Post
Here.

Though if I get enough help, I think we should start creating a Wiki page. Wondering what the HCFR group things of that?
Thanks for the link to the guide! I'd love to see a Wiki page that could include a documented workflow using HCFR that could be updated with tweaks as needed. Great idea.
carillon is offline  
Old 10-14-2014, 10:58 AM
Advanced Member
 
xvfx's Avatar
 
Join Date: Feb 2013
Location: UK
Posts: 807
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 148 Post(s)
Liked: 97
Quote:
Originally Posted by orion2001 View Post
1) Go to preferences->Advanced and set Grayscale dE handling to "Absolute Y w/ Gamma". I haven't used this myself, but this should in theory now take your luminance target into account for the dE calculations. You will find large errors if you now do a sweep and you will need to lower RGB values till you are at D65 AND also meet the required luminance target for your current gray level as determined by your Gamma curve choice in order to minimize dE.
If only CIE2000 stayed for colour because once you start altering defaults it uses the same mode across all. Then starts becoming screwy with other readings.

Switch to CIE2000 and your greyscale and colour use the same just as greyscale and colour uses Absolute Y w/ Gamma.
xvfx is offline  
Old 10-14-2014, 11:07 AM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by xvfx View Post
If only CIE2000 stayed for colour because once you start altering defaults it uses the same mode across all. Then starts becoming screwy with other readings.

Switch to CIE2000 and your greyscale and colour use the same just as greyscale and colour uses Absolute Y w/ Gamma.

I assume that is only when I use an absolute Y w/Gamma. If I want BT.1886, I leave it as is, correct?
pisymbol is offline  
Old 10-14-2014, 11:15 AM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by xvfx View Post
If only CIE2000 stayed for colour because once you start altering defaults it uses the same mode across all. Then starts becoming screwy with other readings.

Switch to CIE2000 and your greyscale and colour use the same just as greyscale and colour uses Absolute Y w/ Gamma.
Quote:
Originally Posted by orion2001 View Post
As such what you are doing is pretty much spot on. The problem is that by default, HCFR calculates dE for grayscale without taking into account Gamma curve target. As a result, when you are making your 2 and 10 pt white balance adjustments, all you are doing are ensuring that the white chromaticity (xy coordinate) is accurate to D65 target (i.e., neutral white). However, minimizing dE does not ensure that you are also matching your luminance target as desired by your reference Gamma curve of choice.
I understand now completely what you are saying. In a nutshell, HCFR doesn't take into Y when I am matching xy for D65 on the CIE graph. So even though my dEs look fantastic, they are still technically off because the Y values are not within spec. Right?

Quote:
In order to also match your target gamma curve correctly, there are a couple of options:

1) Go to preferences->Advanced and set Grayscale dE handling to "Absolute Y w/ Gamma". I haven't used this myself, but this should in theory now take your luminance target into account for the dE calculations. You will find large errors if you now do a sweep and you will need to lower RGB values till you are at D65 AND also meet the required luminance target for your current gray level as determined by your Gamma curve choice in order to minimize dE.
Is this what I want even if I want to target BT.1886 (the defaults)?

NOTE: Currently it's grayed out so I'm not sure how to enable it just yet (though I don't have the iD3 connected).

Quote:
2) The other option is to keep doing as you are doing. Now that you have done a grayscale measurement, you can look at the grayscale table values for each step from 10 - 90%. Then focus on each level and dial in the target luminance. For example: Let's look at your 90% Y value. It was measured to be 113 cd/m2. If you scroll down in the table view (or you can click the "down arrow" in the little widget next to the "Editable Data" checkbox to reveal lower lines in the table) you will see the Target Y also in cd/m2. Let's say this reads 108 cd/m2. This means that your display is outputting too much light at 90 IRE. So now highlight the 90 IRE column, then select free measure and display the 90 IRE pattern on the display. Keep tweaking RGB down till you still have all bars showing 100% (i.e. you have a neutral white) but your measured Y is now lowered to 108 (or close to it). Repeat the process for each IRE step and you will end up with neutral whites and your luminance curve will match your target curve as determined by your reference gamma curve selection.
Should I try to take 21% of my target ftl at neutral white as an approximation for the "Color" setting, or will the above give me what I want? (I guess the answer it depends on the TV).

Quote:
If everything pans out correctly, your gamma curve should look like the gamma curve in my signature (Click the "Spoiler" tag in that post to reveal my full calibration with all the calibration curves)
Sexy.

Quote:
Unfortunately, the luminance curve can be misleading as even big errors in gamma can look very small in that plot. The main curve of interest is the gamma curve which shows how far off the luminance target you are for each step increment in white/gray. As you can see in your curve, your measured gamma is substantially off the target gamma. The average gamma should not be something you should bother about. A lot of poor calibrations by folks are due to them assuming that matching average gamma is good enough. With a 10pt control, there is no reason you shouldn't be able to pretty much nail the target gamma curve.

I'd leave it as is for now.

Good luck!
Please check in from time to time orion, your help has been invaluable and I'd like to document all this.
pisymbol is offline  
Old 10-14-2014, 11:31 AM
Advanced Member
 
xvfx's Avatar
 
Join Date: Feb 2013
Location: UK
Posts: 807
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 148 Post(s)
Liked: 97
Quote:
Originally Posted by pisymbol View Post
I assume that is only when I use an absolute Y w/Gamma. If I want BT.1886, I leave it as is, correct?
No. BT.1886 is on the main tab. Your gamma measurements you want to run thats selected. As the colour you want to measure is on the last tab. You can't set one for greyscale and one for the CIE. Unless you use HCFR's default settings.
xvfx is offline  
Old 10-14-2014, 11:34 AM
AVS Special Member
 
LastButNotLeast's Avatar
 
Join Date: Feb 2007
Location: 08077
Posts: 5,083
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 390 Post(s)
Liked: 325
Quote:
Originally Posted by pisymbol View Post
Here.

Though if I get enough help, I think we should start creating a Wiki page. Wondering what the HCFR group things of that?
I think that would be a great idea. And Kal's older guide is definitely more HCFR-specific, though a Wiki would be able to included elements from his new guide. With his blessing, of course, which would not likely be a problem.

I would keep your personal calibration out of it though, since that's more appropriate for the forum (IMHO). I would love a new guide based on Kal's guides, specifically walking through HCFR.

My vote, for what it's worth.

Michael
|Tch0rT| likes this.

Downloadable FREE demo discs:
http://www.avsforum.com/t/1475769/de...ently-authored 

Did you really need to quote that entire post in your reply?
LastButNotLeast is offline  
Old 10-14-2014, 11:35 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,743
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 162
Quote:
Originally Posted by pisymbol View Post
I understand now completely what you are saying. In a nutshell, HCFR doesn't take into Y when I am matching xy for D65 on the CIE graph. So even though my dEs look fantastic, they are still technically off because the Y values are not within spec. Right?
What Orion said is true ... *however* when calibrating the greyscale, getting to D65 (the x,y part) is the most important task. You really need to do things in two passes: The first to minimize delta-xy (balancing RGB,) the second to fix the Y(luminance and gamma.)

The danger here is that the Y component of dE can mask rather huge errors in chromaticity. IOW, you can get a super low dE number and still have the whitepoint way off into the blue range ... particularly at lower luminance levels.
HDTVChallenged is offline  
Old 10-14-2014, 11:37 AM
AVS Special Member
 
LastButNotLeast's Avatar
 
Join Date: Feb 2007
Location: 08077
Posts: 5,083
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 390 Post(s)
Liked: 325
Quote:
Originally Posted by pisymbol View Post
Please check in from time to time orion, your help has been invaluable and I'd like to document all this.
I think you'd find keeping him AWAY would be a challenge.

Fortunately.

Michael

Downloadable FREE demo discs:
http://www.avsforum.com/t/1475769/de...ently-authored 

Did you really need to quote that entire post in your reply?
LastButNotLeast is offline  
Old 10-14-2014, 11:57 AM
Advanced Member
 
orion2001's Avatar
 
Join Date: Jun 2008
Posts: 728
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 448 Post(s)
Liked: 299
Quote:
Originally Posted by pisymbol View Post
I understand now completely what you are saying. In a nutshell, HCFR doesn't take into Y when I am matching xy for D65 on the CIE graph. So even though my dEs look fantastic, they are still technically off because the Y values are not within spec. Right?
Yes. Just a quick clarification in case it isn't clear, the CIE chart only shows chromaticity co-ordinates (xy) and does not show luminance. Many tend to focus on nailing targets on a CIE chart (both for grayscale and primary/secondary colors) but it is also very important to nail the luminance targets for the grayscale and colors. This is why, the Gamma curve (for grayscale) and the luminance curves (for color saturation sweeps) are very important in addition to CIE charts to visually determine if you have a good calibration.

Quote:
NOTE: Currently it's grayed out so I'm not sure how to enable it just yet (though I don't have the iD3 connected).
Not sure why it is grayed out. Is your color formula set to CIE2000?

Quote:
Should I try to take 21% of my target ftl at neutral white as an approximation for the "Color" setting, or will the above give me what I want? (I guess the answer it depends on the TV).
You lost me here. The 21% recommendation is only for color calibration (CMS). The grayscale and gamma calibration is pretty much decoupled from that. So forget about color for now. Focus on your grayscale and making sure your dE is low (i.e. neutral whites) and that your measured luminance (Y) for all steps are very close to the target Y.

Something I should also clarify since you seem a bit confused/unclear about your target Gamma curve. Gamma is a relationship describing increase in luminance for increase in white level (so essentially the luminance transfer function/curve that is shown in the HCFR luminance graph). When you choose a target Gamma curve (and/or a target Gamma value), you are essentially determining what that target luminance curve should look like. A lot of the target Gamma curves such as BT.1886 also use the measured black level during the grayscale measurements to create an offset at the lower end of the curve (this is primarily to prevent blacks from being crushed on displays with high black levels). In either case, once the black and 100% white measurements have been made, they can be used to determine the exact target Y values for your target Gamma curve. This is how HCFR calculates the Y Target listing in the table, and this is also why it won't provide the target Y until you first do a sweep so it has measurements for black and 100% white.

For exactly this same reason, you want to first ensure that your 0% and more importantly 100% white is exactly where you want it to be (i.e., neutral and providing your the desired max luminance based on your viewing conditions) before you spend all this effort dialing in the gamma curve. Making a tweak later on for the 100% point will essentially change all your Y Targets for your target gamma curve and may require tweaking across all the 10points again to meet your target gamma curve.

Quote:
Please check in from time to time orion, your help has been invaluable and I'd like to document all this.
You're very welcome. Just don't take my word for gospel as I consider myself a noob too . I've just done a lot of reading and have also learned a lot from all the great wealth of knowledge and experience from some of the more experienced calibrators and software authors who frequent these forums.

Quote:
Originally Posted by HDTVChallenged View Post
The danger here is that the Y component of dE can mask rather huge errors in chromaticity. IOW, you can get a super low dE number and still have the whitepoint way off into the blue range ... particularly at lower luminance levels.
This is a very good point and I suspect why the default is to have the dE formula only account for chromaticity (xy coordinate) to ensure that you achieve neutral white (which is most important).

Taking all of this into account, I think the easier way to go about things is to try the way I listed (option 2) to hone in on the target luminance values. There probably might be be better ways to do this in HCFR. I'm still learning my way around some of the hidden features/quirks!

-------------------------------------------------------------------------
Recommended calibration settings for Samsung PN60F5300B
orion2001 is offline  
Old 10-14-2014, 12:45 PM
Advanced Member
 
xvfx's Avatar
 
Join Date: Feb 2013
Location: UK
Posts: 807
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 148 Post(s)
Liked: 97
Can anyone remember the discussion that someone used a microphone stand for their iD3? and if so which it was?
xvfx is offline  
Old 10-14-2014, 03:20 PM
AVS Special Member
 
LastButNotLeast's Avatar
 
Join Date: Feb 2007
Location: 08077
Posts: 5,083
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 390 Post(s)
Liked: 325
I use a tripod. The D3 screws right in to the plate that normally holds a camera. Works well, and I had one hanging around!

Downloadable FREE demo discs:
http://www.avsforum.com/t/1475769/de...ently-authored 

Did you really need to quote that entire post in your reply?
LastButNotLeast is offline  
Old 10-14-2014, 03:49 PM
Member
 
Join Date: Sep 2014
Posts: 15
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 0
Quote:
Originally Posted by pisymbol View Post

Should I try to take 21% of my target ftl at neutral white as an approximation for the "Color" setting, or will the above give me what I want? (I guess the answer it depends on the TV).
Here are the steps I find was successful and quick:
1. Download HCFR calibration disk
2. Watch the HDTV Calibration video on it. Warms up the set.
3. Run “Basic Setting” calibration; Contrast, Brightness, color and Sharpness. Pay special attention to any color shifts in setting contract too high. This is a sign of clipping.
4. Run Misc. Patterns, Pattern A (additional) and look at Pattern #3 Color Steps, you should see a step for each color. Make sure you see all shades of each color. If any color at 235 and the one next to it are the same you need to turn down contrast some more because you are clipping. Look at Pattern #4 also, color clipping.
5. Run HCFR.
6. Set the gamma you want in preferences. You need to so you can dial it in latter. (power, 2.2, 2.4 ....). Make sure set Measured Gamma is not selected. (see 1st Thumbnail)
7. Use the Blu-ray disk for patterns since there are so many things that can go bad by using a PC (did you set the full range or limited on the video card?, did you change the output to match in the HCFR preferences 0-255 or 16-255...) plus it is easier to nail down gama when using the disk because you can do continuous measurements ( Make setting 10 point and gamma a breeze). Hit Active integration (number 1 on second thumbnail) 3 times to make sure it is off. Looks like a bug and shows off when you start but it is on. Hitting it once states it is off, twice it is on, third time it is off.

EDIT: Zoyd said there are several editions of ver 3.1.6 and the latest version appears to have corrected this issue with active integration. (I wish it was labeled version 3.1.6.3 or 3.1.6c or even 3.1.7 or something vs having 3 versions with the same version number).

8. Run the 100% gray pattern at 100% and set backlight to your desired brightness (do not use contrast to set your Y, you set that already and you may cause clipping.) Use continuous measuring (number 2 the large green arrow on second thumbnail) and to dial in your backlight level.
9. Adjust you 100% RGB levels while you at it (go down so you don't run into clipping issue if you go up). I set my contrast 1 click lower so I have headroom.
10. Turn off continuous reading (for now)
11. Run a gray level sweep.
12. Check your Gamma chart.
13. Now you dial in your gamma, look at the results from you grayscale. The "Target" Y (number 3 on second thumbnail) is hidden (not really, you have to scroll down so it shows up, they should have made the box bigger so that shows up without scrolling down)
14. Now go to 90% on the DVD, turn on continuous monitoring(number 2 big green arrow) and dial in the Y target for 90% by adjusting the 10 point White Balance RGB at the 90%. Look at the levels and balance the RGB and Y (number 4 thumbnail 2) to target level (number 3 thumbnail2) (do not re-adjust 100% after you set the level or you will mess up all the other levels by changing targets. You can clean it up by slight tweek during your second or third sweep.)
15. Go to 80% on Blu-Rayand adjust the 80% white level to the Target Y....
16. Check Brightness pattern, contrast and clipping, Step 3
17. Do all and rerun Gray sweep and your Gamma should be real close to target line. Repeat 6-10 until your happy, bored or run out of beer.
Attached Thumbnails
Click image for larger version

Name:	1.jpg
Views:	34
Size:	54.5 KB
ID:	311426   Click image for larger version

Name:	2.jpg
Views:	36
Size:	84.7 KB
ID:	311458  

Last edited by AMS67067; 10-18-2014 at 01:38 PM. Reason: JPG 2 added, changed to 235, corrections
AMS67067 is offline  
Old 10-14-2014, 04:40 PM
Member
 
Join Date: Sep 2014
Posts: 15
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 0
After just 1 run you should have a somewhat flat gamma (see thumbnail). Just redo 1 or two more time and it should be nice and flat or to whatever gamma you selected in preference. The target values adjust to whatever gamma you select.
Attached Thumbnails
Click image for larger version

Name:	3.jpg
Views:	28
Size:	75.8 KB
ID:	311506  

Last edited by AMS67067; 10-15-2014 at 07:47 PM.
AMS67067 is offline  
Old 10-14-2014, 05:05 PM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by xvfx View Post
No. BT.1886 is on the main tab. Your gamma measurements you want to run thats selected. As the colour you want to measure is on the last tab. You can't set one for greyscale and one for the CIE. Unless you use HCFR's default settings.

You and orion aren't being clear! :-)

I understand pretty much everything that has been said EXCEPT how it translates to actual GUI widgets on HCFR.

For the References tab, I have REC709 75%/75% select as well as ITU-R BT.1886 as my gamma transfer function (that I assume selects my EOTF, i.e. the math). I have a Reference Gamma of 2.22 which is grayed out, not changeable. I was PRETTY confident that is fine up until about a few posts back.

In the Advanced Tab, I have under Delta E, Color Difference Formula set to Recomended. Why would I change this to CIE2000? My Gray Scale dE handling is grayed out as well as my white reference wighting (even if I change to CIE2000 that remains true). I am not even sure how to enable that.

What I am not understanding right now is A) Why would I choose CIE2000 vs Recommended? and B) Why is the target gamma of 2.22 under the default BT.1886 not what I want if I want to calibration for 1886?

The rest I think I got. I need to now go back and match Y values to my reference which is determined by my gamma function and my grayscale sweeps.

Please correct me if the above is wrong or unclear.

I understand the CIE chart. I understand it doesn't express the luminance targets. What I don't understand right now is HCFR settings and they correlate with measurements! :-)
pisymbol is offline  
Old 10-14-2014, 05:10 PM
Senior Member
 
Join Date: Sep 2014
Posts: 274
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 28
Quote:
Originally Posted by AMS67067 View Post
Here are the steps I find was successful and quick:
6. Set the gamma you want in preferences. You need to so you can dial it in latter. (power, 2.2, 2.4 ....). Make sure set Measured Gamma is not selected. (see 1st Thumbnail)
I only uncheck BT.1886. That is what is royally confusing me. I want my gamma function to adhere to BT.1886 function so I'm at a lost now!

Help! :-)
pisymbol is offline  
Old 10-14-2014, 05:22 PM
Member
 
Join Date: Sep 2014
Posts: 15
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 0
Quote:
Originally Posted by pisymbol View Post
You and orion aren't being clear! :-)

What I am not understanding right now is A) Why would I choose CIE2000 vs Recommended? and B) Why is the target gamma of 2.22 under the default BT.1886 not what I want if I want to calibration for 1886?

T)
A. Default is CIE76 (i am pretty sure) and it is the older standard before HDTV . You should use CIE 94 or 2000 just because that is what with HDTV. Has stricter Delta E . google it and read up .

EDIT: Recommended uses CIE76 for Grayscale andCIE2000 for Color. It uses the best of both worlds. I guess recommended is recommended

B. Make it anything you want, select ITU-R BT 1886 in Preference Reference. This is commonly used in a really dark room with low light output setting. If you have a FALD TV set using 1886 will default to flat gamma 2.4 since it is function of how much light you output at 0 %. It ramps up faster at lower light stimulus.

Last edited by AMS67067; 10-18-2014 at 01:29 PM. Reason: Correction
AMS67067 is offline  
 
Thread Tools


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