Display Calibration Calculator - Page 2 - AVS Forum
Forum Jump: 
 
Thread Tools
Old 06-22-2008, 01:39 PM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by kjgarrison View Post

Run5 ColorHCFR is the screen showing ColorHCFR dEs. In the comments, you can see the settings used to move the green primary closer to it's spot, which because of the angle to red becomes slightly undersaturated. Also, of course, the setting to get hue/saturation/luminance of all colors to good dE values per Accupel.

Kj, your use of the calculator is spot on. Looking at your CIE chart, I'm curious, are you able to move green more toward red at all so it falls closer to the line from the ideal green to D65? It looks like you've ticked the Red component of green up to 40. Also, the same for the red point, can you add blue to it to get it closer to the line between red and D65? Or does this throw things off too much?

cheers,


--tom
thomasl is offline  
Sponsored Links
Advertisement
 
Old 06-22-2008, 02:20 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by thomasl View Post

Kj, your use of the calculator is spot on. Looking at your CIE chart, I'm curious, are you able to move green more toward red at all so it falls closer to the line from the ideal green to D65? It looks like you've ticked the Red component of green up to 40. Also, the same for the red point, can you add blue to it to get it closer to the line between red and D65? Or does this throw things off too much?

cheers,


--tom

Tom, yes I could do both. Should I?

The line from D65 to CIE green is pretty close to where it is, though. It would undersaturate green only a little more. The sliders go quite a bit higher, maybe 90. Can't recall exactly, but I know it's more than 80.

Red could be pulled towards blue with the blue slider. The dotted line going across through D65 to cyan is where it would intersect and that looks like an x of 0.6. Since y would be perfect, who knows, maybe red primary dE would be better.

As you can see, I have all the secondaries on the line from their primary colors that passes through D65. When I measure saturation with ColorHCFR the dots track nicely from D65 to where the secondary is. Perhaps your idea would be better. However since the lines between the primaries would not be proper CIE, the secondaries might be along the "proper" CIE track, but they would not be at the right ending point

Should I test your idea?
kjgarrison is offline  
Old 06-22-2008, 02:55 PM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by kjgarrison View Post

As you can see, I have all the secondaries on the line from their primary colors that passes through D65. When I measure saturation with ColorHCFR the dots track nicely from D65 to where the secondary is. Perhaps your idea would be better. However since the lines between the primaries would not be proper CIE, the secondaries might be along the "proper" CIE track, but they would not be at the right ending point

You're correct if adding blue to red simply moves it in or down along the line between red and blue. It'll just shrink the gamut more and this will cause the optimal magenta to shrink in a bit as well - thus not at the right ending point. It might be worth a try just to see what happens as you add a little blue to red. As for green, it looked like if it just slid over a bit along the x axis, it would line up perfectly but be undersaturated. But now that I look at the x,y values again, x is very close as it is and probably within the variance of your colorimeter if it's an Eye One. I wouldn't tinker with green at all. Overall, your dE values using HD709 as the reference are very good. Everything I've read says a dE < 10 is good for dELUV when measuring primary/secondary color error. So, red is the only one a bit off because x,y is a tad off and it looks like the luminance value for your custom gamut is about ~.23 as opposed to ~.21 for the HD709 red. Of course, the most important question is how do things look?

cheers,


--tom
thomasl is offline  
Old 06-22-2008, 03:40 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by kjgarrison View Post

Interestingly, gregr came over to our discussion and seemed to agree with 'the other guy', which left me wondering what the Accupel calculator is useful for.

I've answered this before but I'll try again.

1. The FIRST purpose of the AccuPel Display Calibration calculator is to adjust the display for STANDARD YCbCr/YPbPr to RGB Color Decoding. Most front projectors will have standard YCbCr/YPbPr to RGB color decoders, but Tint (Hue) and Color controls modify the color decoding. (Many non-front projectors have non-standard color decoders, but may include as many as 6 additional specific color decoder adjustments in a service menu.)

In this case you should enter your measured primaries and gray reference in the Target Values at the top of the Color Gamut screen and press Compute. The calculator will compute the Y values of the primary and the x,y,Y values of the complementary colors with standard color decoding. You can enter your measured primary and complementary color values at the bottom as you adjust the color decoding and the calculator will compute the dE (CIELUV 1976 - my preference) deviation from the Target values each time you press the Compute button.

2. The SECOND purpose of the AccuPel Display Calibration calculator is to adjust the display's total colorimetry to match a Target display colorimetry "as closely as possible". If the display (or an external processor) has a fully functional CMS system, then I recommend the display's Color Decoder first be adjusted for Standard color decoding (#1 above), and then the CMS processing be adjusted to match the Target display colorimetry. This is the ideal/preferred case. However, many displays have no CMS or an incomplete CMS, and in that case it may be necessary to adjust whatever controls are provided (partial CMS and/or Color Decoding) to match the Target display colorimetry "as closely as possible".

You must decide what Target display colorimetry you want to try to match. You may choose a Target display colorimetry based on a set of Standard primaries (Rec. 709, SMPTE C, or EBU) or you may enter a set of Custom/Measured primaries and a gray reference. If you enter Custom/Measured primaries or a non-D65 gray reference and then press the Compute button, the calculator will compute Target x,y,Y values based on Standard Color Decoding. If you enter measured x,y,Y values of the primary and complementary colors at the bottom, the dE (CIELUV 1976) deviation from the Target values you chose at the top will also be calculated when you click the Compute button.

3. The THIRD purpose of the AccuPel Display Calibration calculator is to adjust the display's Grayscale tracking as closely as possible to D65. You can enter the x.y values of the grayscale and the calculator will compute the dE (CIELUV 1976) values of the grayscale.

4. The FOURTH purpose of the AccuPel Display Calibration calculator is to adjust the display's Gamma curve to a Target Gamma curve. You may select any Gamma value from 1.80 to 2.20 in 0.05 increments and the calculator will compute the Target Y values from 5% to 90% based on your measured 100% Y value. It will also compute the actual incremental Gamma values for measured values that you enter. You can also enter a black level (0%) or a Full-Field Contrast Ratio to compute the Gamma values relative to a non-zero black level.

The latest version of the AccuPel Display Calibration Calculator is version 1.23. The only differences from previous versions are the labeling of the SMPTE C and EBU standard primary buttons, and a 4th decimal place was added to the computed color gamut Y Target values because SMPTE standards also use 4 decimal places.

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 06-22-2008, 06:23 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Thank you Greg.

The way you describe is how I think I have been using the calculator. Just the way it is laid out suggests how to use it, actually. Maybe I'm missing something critical, but as I read your descriptions of 1 and 2, I kept thinking .. yes, that's what I thought and what I do.

For some reason when you came over to the Samsung 71 calibration thread, it seemed to those of us involved in the debate that you were saying something different about how to use it. Maybe you were saying something else about what the calculator's modified Y values for primaries and xyY values for secondaries signifies. You may recall that the debate was whether to place magenta where the Accupel calculator says it should go, or place it at the Rec709 magenta spot. The Samsung 71s have green off to the left, and the definition of magenta as the location of the intersect of a line drawn from green through D65 and a line drawn from blue to red, causes magenta to be substantially closer to red than the Rec709 spot. The calculator shows lower dE with this new magenta spot.

So, if I read what you have written here correctly, you are saying to put the secondaries where the calculator says to put them even if that is not the same xy coordinates as they would occupy in a perfect Rec709 CIE chart.

Thank you very much for this detailed description of "How to". You should consider copying and pasting this in your first post in case other noobs meander into this thread.
kjgarrison is offline  
Old 06-22-2008, 06:28 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by thomasl View Post

You're correct if adding blue to red simply moves it in or down along the line between red and blue. It'll just shrink the gamut more and this will cause the optimal magenta to shrink in a bit as well - thus not at the right ending point. It might be worth a try just to see what happens as you add a little blue to red. As for green, it looked like if it just slid over a bit along the x axis, it would line up perfectly but be undersaturated. But now that I look at the x,y values again, x is very close as it is and probably within the variance of your colorimeter if it's an Eye One. I wouldn't tinker with green at all. Overall, your dE values using HD709 as the reference are very good. Everything I've read says a dE < 10 is good for dELUV when measuring primary/secondary color error. So, red is the only one a bit off because x,y is a tad off and it looks like the luminance value for your custom gamut is about ~.23 as opposed to ~.21 for the HD709 red. Of course, the most important question is how do things look?

cheers,


--tom

It looks like red is more off because I have already dialed a bunch of red into green. If I left green at default (with no red in it), the dE for green would be ~15%. Luckily green is outside the triangle (oversaturated), and at least some of green's error can be reduced.

Red is undersaturated and nothing can be done unless there is something in the service menu, which I do not have confidence in mucking around in.
kjgarrison is offline  
Old 06-22-2008, 06:40 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
I forgot I had these measurements on a thumb drive.

This is the CIE chart with default settings.

Red is red50 green0 blue0

Green is red0 green50 blue0

Blue is red0 green0 blue50

Yellow red50 green50 blue0

Cyan red0 green50 blue50

Magenta red50 green0 blue50
LL
kjgarrison is offline  
Old 06-22-2008, 08:31 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by kjgarrison View Post

Thank you Greg.
So, if I read what you have written here correctly, you are saying to put the secondaries where the calculator says to put them even if that is not the same xy coordinates as they would occupy in a perfect Rec709 CIE chart.

I still don't think you are understanding what I am trying to convey (may be my fault). You have to decide what set of primary and complementary Target colors you want to match. The calculator can not select that set for you. It doesn't know what you like or don't like, or what you believe about color accuracy or don't believe. It is preprogrammed with 3 standard choices (Rec 709, SMPTE C, EBU) that you may wish to use (I would). OR you can choose to use any Custom set that you want by just entering the primary targets and gray reference yourself. If you do that it will calculate the complementary color targets for you. So it is up to you to pick the target colorimetry that you want to use. Then it will help make adjustments by telling you the dE deviations of your measured/adjusted values from the Target set you chose to use.

If I am reading your posts correctly (maybe I'm not) you are choosing to use the display's native primaries to produce your final Target display colorimetry. I wouldn't do that, but that is up to you. (I would only use the display's native primaries to be sure (adjust) that the color decoding was standard, then I would use a standard set of primaries for the display's final Target colorimetry.)

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 06-22-2008, 11:04 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by gregr View Post

I still don't think you are understanding what I am trying to convey (may be my fault). You have to decide what set of primary and complementary Target colors you want to match. The calculator can not select that set for you. It doesn't know what you like or don't like, or what you believe about color accuracy or don't believe. It is preprogrammed with 3 standard choices (Rec 709, SMPTE C, EBU) that you may wish to use (I would). OR you can choose to use any Custom set that you want by just entering the primary targets and gray reference yourself. If you do that it will calculate the complementary color targets for you. So it is up to you to pick the target colorimetry that you want to use. Then it will help make adjustments by telling you the dE deviations of your measured/adjusted values from the Target set you chose to use.

If I am reading your posts correctly (maybe I'm not) you are choosing to use the display's native primaries to produce your final Target display colorimetry. I wouldn't do that, but that is up to you. (I would only use the display's native primaries to be sure (adjust) that the color decoding was standard, then I would use a standard set of primaries for the display's final Target colorimetry.)

Thank you for your patience. It's not you, it's me.

You said (above post) ...

Many non-front projectors have non-standard color decoders, but may include as many as 6 additional specific color decoder adjustments in a service menu.)

In this case you should enter your measured primaries and gray reference in the Target Values at the top of the Color Gamut screen and press Compute. The calculator will compute the Y values of the primary and the x,y,Y values of the complementary colors with standard color decoding.

To me this sounds like instructions on what to do with "non-standard" color.

OK, so I choose Rec709.

Then I put my results in and see my primary colors' dE values.

Then I adjust my primary colors to get as close to Rec709 as I can get them, but I can't get them perfect. So, no, I am not choosing to use the display's native primaries to produce my final target, but I am stuck with sub-optimal primaries. You could call it a Custom set, but it isn't by choice. (With the 71 I was stuck with the native primaries and they were waaay off.)

So then I say, well OK if these are the non Rec709 primaries that I have to live with, what should I do with their Y values and what should I do with the xyY values of the secondaries? I have been doing what you seem to have said above to do.

I have been using the Copy function of your calculator to put my actual, improved (not native, but as close to my chosen Rec709 as is possible) primaries' xy values on top. A Custom set as you called it. Then I hit calculate and see what changes the calculator says to make with the primaries' Y values and the secondaries' xyY values. Then I tweak primary Y and secondary xyY until I get as close to the targets as indicated by low dE's.

Are you saying not to do that, but instead just live with the significantly "off" primaries' xy and get the primaries' Y and secondaries' xyY as close to Rec709 (the standard of my choosing) targets as possible?
kjgarrison is offline  
Old 06-23-2008, 11:50 AM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by kjgarrison View Post

Red is undersaturated and nothing can be done unless there is something in the service menu, which I do not have confidence in mucking around in.

Red looks to be pretty much in the same position between your before and after CIE charts - which makes sense. It's leaning toward green a bit and is under-saturated but there is no way to fix those. One way to think about it is, there is no way to make red "redder/more saturated" if that red is the native red of the display.

cheers,


--tom
thomasl is offline  
Old 06-23-2008, 12:58 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by kjgarrison View Post

Thank you for your patience. It's not you, it's me.

You said (above post) ...

Many non-front projectors have non-standard color decoders, but may include as many as 6 additional specific color decoder adjustments in a service menu.)

In this case you should enter your measured primaries and gray reference in the Target Values at the top of the Color Gamut screen and press Compute. The calculator will compute the Y values of the primary and the x,y,Y values of the complementary colors with standard color decoding.

To me this sounds like instructions on what to do with "non-standard" color.

Yes, to adjust/verify that a color decoder performs standard color decoding you need to use the "non-standard" native primaries to generate the Target values.

Quote:
Originally Posted by kjgarrison View Post

OK, so I choose Rec709.

Then I put my results in and see my primary colors' dE values.

Then I adjust my primary colors to get as close to Rec709 as I can get them, but I can't get them perfect. So, no, I am not choosing to use the display's native primaries to produce my final target, but I am stuck with sub-optimal primaries. You could call it a Custom set, but it isn't by choice. (With the 71 I was stuck with the native primaries and they were waaay off.)

OK, then you are using the Rec 709 Targets at this point in your description. You are not using a Custom set of Targets based on the native primaries at this point in what you have described.

Quote:
Originally Posted by kjgarrison View Post

So then I say, well OK if these are the non Rec709 primaries that I have to live with, what should I do with their Y values and what should I do with the xyY values of the secondaries? I have been doing what you seem to have said above to do.

I have been using the Copy function of your calculator to put my actual, improved (not native, but as close to my chosen Rec709 as is possible) primaries' xy values on top. A Custom set as you called it. Then I hit calculate and see what changes the calculator says to make with the primaries' Y values and the secondaries' xyY values. Then I tweak primary Y and secondary xyY until I get as close to the targets as indicated by low dE's.

At this point you have changed from using the Rec 709 Targets to Targets based on your own native display primaries.

Quote:
Originally Posted by kjgarrison View Post

Are you saying not to do that, but instead just live with the significantly "off" primaries' xy and get the primaries' Y and secondaries' xyY as close to Rec709 (the standard of my choosing) targets as possible?

If you want to match one of the standard colorimetries (Rec 709, SMPTE C, or EBU) as closely as possible (which may not be very close for your display depending on what adjustments it has) then you need to use those as the Target values and not the native primaries that you copied into the Target section above.

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 06-23-2008, 01:34 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by gregr View Post

If you want to match one of the standard colorimetries (Rec 709, SMPTE C, or EBU) as closely as possible (which may not be very close for your display depending on what adjustments it has) then you need to use those as the Target values and not the native primaries that you copied into the Target section above.

Thank you gregr.

Before I let this completely go, there is one nagging question left unanswered.

My previous display was a Samsung 71 and it had zero ability to adjust primaries. Not xy and not Y separately from what could be done, of course, with all colors in the main color control.

When I input all the xyY data as I described above and made this 'custom' or 'native' color gamut, the Y values for all colors were natively almost exactly where the Accupel calculator said they should be. This, more than anything else made me think that when dealing with a non-spec color gamut, this is, if not the best way to use the calculator, at least an OK way. It was almost like the Samsung engineers used the same logic/calculations to set the 71's up as the Accupel calculator.

Anyway I really do appreciate your help Greg.
kjgarrison is offline  
Old 06-23-2008, 01:57 PM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by gregr View Post

1. The FIRST purpose of the AccuPel Display Calibration calculator is to adjust the display for STANDARD YCbCr/YPbPr to RGB Color Decoding. Most front projectors will have standard YCbCr/YPbPr to RGB color decoders, but Tint (Hue) and Color controls modify the color decoding. (Many non-front projectors have non-standard color decoders, but may include as many as 6 additional specific color decoder adjustments in a service menu.)

Greg, if one's display has CMS controls, are you saying that you'd first adjust color/tint using the display's native primaries/reference white point (fine tune the color decoder) - and then leave those two settings alone and go into the CMS controls in order to then use those controls to try and match a target such as HD709 or SMPTE-C? And what if one only has Color/Tint to work with, do you just recommend using them to match a target (e.g. HD709, SMPTE-C) and not use them as you've described in #1?

many thanks,


--tom
thomasl is offline  
Old 06-23-2008, 02:47 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by kjgarrison View Post

My previous display was a Samsung 71 and it had zero ability to adjust primaries. Not xy and not Y separately from what could be done, of course, with all colors in the main color control.

When I input all the xyY data as I described above and made this 'custom' or 'native' color gamut, the Y values for all colors were natively almost exactly where the Accupel calculator said they should be.

That verifies/demonstrates that it has standard color decoding.

Quote:
Originally Posted by kjgarrison View Post

This, more than anything else made me think that when dealing with a non-spec color gamut, this is, if not the best way to use the calculator, at least an OK way. It was almost like the Samsung engineers used the same logic/calculations to set the 71's up as the Accupel calculator.

Any display with standard color decoding should work that way.

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 06-23-2008, 03:24 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by thomasl View Post

Greg, if one's display has CMS controls, are you saying that you'd first adjust color/tint using the display's native primaries/reference white point (fine tune the color decoder) - and then leave those two settings alone and go into the CMS controls in order to then use those controls to try and match a target such as HD709 or SMPTE-C?

Yes, the first thing I would do is confirm/adjust the color decoding to be standard if possible. For one reason, there are some sources and some displays that use incorrect color decoding matrices (design errors). They mix up the standard-definition (Rec 601) and high-definition (Rec 709) matrices. So it's a good idea to first confirm that the correct matrices are being used. (If they should happen to be wrong there are often manual settings that can be changed to use the right matrix, and sometimes they are wrong because someone incorrectly set one of these manual selections.) Secondly, in non-front projectors the decoding matrix is sometimes very non-standard (by design - for bad reasons) but there may be controls in a service menu to adjust it to standard performance - although those controls are often difficult for an inexperienced person to understand how to use. Third, the simple Color/Hue controls may be incorrectly set in a service sub-menu and the default setting in a user menu may not be correct (that isn't too common except in the second case above). Most important to remember, if the color decoding is not standard, the color accuracy will be wrong even if you adjust a CMS to match standard primary and complementary colors.

Quote:
Originally Posted by thomasl View Post

And what if one only has Color/Tint to work with, do you just recommend using them to match a target (e.g. HD709, SMPTE-C) and not use them as you've described in #1?

If there are no CMS controls, only Color/Hue controls, then in most cases the only thing you can do to improve the color matching to a standard is to adjust the color decoding (Color/Hue) controls and the reference white point. But its still a good idea to first check the color decoding as I described in #1 (and above) to make sure that you don't have a correctable matrix mismatch problem. But after you have eliminated those possibilities you are basically left with only the Color/Hue controls (or other color decoder controls that might be in a service menu - usually only found in non-front projectors) and the reference white point. You can move the reference white point "some" to improve the overall color accuracy and your eyes will adapt to a non-D65 white reference, but if you push it too far you will notice the difference especially on black and white source material.

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 06-24-2008, 05:00 AM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by gregr View Post

(that isn't too common except in the second case above). Most important to remember, if the color decoding is not standard, the color accuracy will be wrong even if you adjust a CMS to match standard primary and complementary colors.

Greg, thanks for the explanation. Please bear with me if this seems like a stupid question but if a display (and assuming the source chain is ok as well) has standard/correct color decoding, is it true that if a CMS is correctly implemented, all one would need to be able to do is shift/move the primary colors to their target (e.g. HD709) positions (and adjust the white point to D65) and the primary Y and secondary xyY should all be on target after one does this?

many thanks again,


--tom
thomasl is offline  
Old 06-24-2008, 01:22 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by thomasl View Post

Greg, thanks for the explanation. Please bear with me if this seems like a stupid question but if a display (and assuming the source chain is ok as well) has standard/correct color decoding, is it true that if a CMS is correctly implemented, all one would need to be able to do is shift/move the primary colors to their target (e.g. HD709) positions (and adjust the white point to D65) and the primary Y and secondary xyY should all be on target after one does this?

many thanks again,


--tom

If the CMS is implemented using a linear 3x3 RGB to RGB matrix conversion from the native primaries to the "corrected" pseudo primaries, then that is true because setting the white reference point to D65 correctly sets the primary Y values, and the complementary xyY values are a linear combination of the corrected RGB values. However, if the CMS is done using LUTs (Look Up Tables), or non-linear transforms, then anything can happen.

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 06-24-2008, 02:39 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by gregr View Post

If the CMS is implemented using a linear 3x3 RGB to RGB matrix conversion from the native primaries to the "corrected" pseudo primaries, then that is true because setting the white reference point to D65 correctly sets the primary Y values, and the complementary xyY values are a linear combination of the corrected RGB values. However, if the CMS is done using LUTs (Look Up Tables), or non-linear transforms, then anything can happen.

Where is the best place to learn how to do that?

Something like a Matrix Conversion For Dummies.
kjgarrison is offline  
Old 06-24-2008, 08:32 PM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by gregr View Post

If the CMS is implemented using a linear 3x3 RGB to RGB matrix conversion from the native primaries to the "corrected" pseudo primaries, then that is true because setting the white reference point to D65 correctly sets the primary Y values, and the complementary xyY values are a linear combination of the corrected RGB values. However, if the CMS is done using LUTs (Look Up Tables), or non-linear transforms, then anything can happen.


Thanks Greg - that makes sense to me (finally). As a software developer/computer person, I'm intrigued by how a lot of this stuff is actually implemented.

thanks again for the continuing education,


--tom
thomasl is offline  
Old 04-22-2009, 05:04 PM
Senior Member
 
dave999z's Avatar
 
Join Date: Jan 2009
Posts: 216
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 16
Sorry for the necro bump...

Quote:
Originally Posted by kjgarrison View Post

I have been using the Copy function of your calculator to put my actual, improved (not native, but as close to my chosen Rec709 as is possible) primaries' xy values on top. A Custom set as you called it. Then I hit calculate and see what changes the calculator says to make with the primaries' Y values and the secondaries' xyY values. Then I tweak primary Y and secondary xyY until I get as close to the targets as indicated by low dE's.

This is a great explanation. This is exactly what I would think to do in order to get the best result you're going to get from your particular set's limitations. For example, my green primary's xy just cannot get to the Rec709 point. So I have a choice, either (1) calibrate the xyY of every other primary and secondary to get as close to the Rec709 points as possible (though in the end green and yellow will be slightly off), or (2) use the calculator to generate new primary Y and secondary xyY values and then calibrate to those values. It seems to me that #2 would result in skin tones that look more like real skin tones than #1.

Important side question: if I go route #2, do I just tweak the all the primary Y and all the secondary xyY without recalculating along the way? Or do I tweak red Y, copy it up and recalculate, tweak blue Y, copy it up and recalcuclate, etc?

Quote:
Originally Posted by kjgarrison View Post

Are you saying not to do that, but instead just live with the significantly "off" primaries' xy and get the primaries' Y and secondaries' xyY as close to Rec709 (the standard of my choosing) targets as possible?

I unfortunately do not really understand greg's answer to this question. I do not understand why it is better to aim everything for the Rec709 points when you know that one primary and one secondary can't get there. It seems that would result in a distorted triangle. And what's the use of the calculator then (if Rec709 is your goal)?

Many thanks in advance for any assistance anyone can provide.
dave999z is offline  
Old 04-22-2009, 06:07 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
dave999z, I never did really understand gregr's explanation either. I felt that he was annoyed by my lack of knowledge on the subject, so I stopped asking. He may have answered my questions well, for all I know, but he just didn't understand that he needed to dumb it down. A lot.

Then I found the thread below, which involves a spreadsheet using RGB/XYZ matrix conversion starting simply with RGB primary xy coordinates and going through an XYZ matrix (whatever THAT is) and ending up with desired RGB primary Y values as well as YCM xyY values. It also takes into account (one of the inputs) the measured gamma. I made a few mods to the spreadsheet which converts percentage of white Y for RGBYCM Y values to actual target Y values based on actual white Y values. I then put it all in one place along with desired xy coordinates for YCM.

Fortunately my Samsung LCD has enough color controls to allow these changes, for the most part. The result: I'm very happy with the color I have achieved.

If I recall correctly, I think they said that the Accupel calculator comes up with the same, or close to the same, results. Go figure.

Here's the link for this little thread that has helped me immeasurably. It took me half an hour to find it. I can't believe I didn't subscribe to it!

http://www.avsforum.com/avs-vb/showthread.php?t=1111066
kjgarrison is offline  
Old 04-22-2009, 06:51 PM
Senior Member
 
dave999z's Avatar
 
Join Date: Jan 2009
Posts: 216
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 16
wow, i'm not sure i can really follow all those spreadsheets on there.

i went ahead and just calibrated my primary xy as close to 709 as possible, then used the Accupel calculator to compute the rest of the primary Y and secondary xyY targets, then finished calibrating to those. and i'm pretty pleased with the result, but have to live with it for awhile.


skin tones are really difficult on these samsungs. it's hard not to be too yellow or red, and not to have green shadows.
dave999z is offline  
Old 04-22-2009, 08:48 PM
Advanced Member
 
thomasl's Avatar
 
Join Date: Apr 2004
Posts: 592
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by dave999z View Post

I unfortunately do not really understand greg's answer to this question. I do not understand why it is better to aim everything for the Rec709 points when you know that one primary and one secondary can't get there. It seems that would result in a distorted triangle. And what's the use of the calculator then (if Rec709 is your goal)?

Dave, there are two "camps" here I believe - one says to adjust your primaries luminance (Y) and secondaries (xyY) based on your white point and primaries xy info - think of this as your "custom" gamut. The other "camp" says you should stick with trying to target a known standard even if your primaries don't line up with the standard. I don't believe that there is right or wrong answer.

As for the uses of the calculator, Greg did a nice job summarizing a few posts up what the calculator can be used for. If you do use it to target your custom display gamut, then all you need to do is enter your primary locations and white point and then run the calculation. You don't need to re-do any calculations unless the locations of your primaries or your white point change.

hope this helps,


--tom
thomasl is offline  
Old 04-22-2009, 10:35 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
I have tried both methods many times. Well, not many times for the "matrix" method. For that method it looked great the first time, and for months now I have not felt the urge to tinker with anything to do with color.

Here is the modified spreasheet I use.

Just input the xy values for RGB in columns GHI, rows 9-10; input white xyY in J9-11; and input gamma in B5. Target Y values for RGBYCM are in row 13, G,H,I,K,L,M.

I put the Y values for RGB also, but that is just to see how much adjustment is needed. It doesn't effect the calculations.

 

SaturationReferencesMOD.zip 22.033203125k . file
kjgarrison is offline  
Old 04-23-2009, 10:58 AM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by kjgarrison View Post

I never did really understand gregr's explanation either. I felt that he was annoyed by my lack of knowledge on the subject, so I stopped asking. He may have answered my questions well, for all I know, but he just didn't understand that he needed to dumb it down. A lot.

I'm sorry if anyone got the impression that I was annoyed last June 2008 by the questions being asked. I've literally answered hundreds (maybe a thousand or more) questions on this forum over the years and I try not to get annoyed with anyone asking a question unless they persist in arguing for some point they don't understand after it's been explained to them multiple times by myself and others. That wasn't the case in this thread, so I don't remember being annoyed with anyone.

The best explanation I can give for the purpose of the calculator is in one of my replies above where I enumerated 4 uses of the calculator. If that confused anyone just ask a specific question about that posting and I will try to clarify. I'm not sure where the explanations need to be "dumbed down - a lot" to make them more understandable.

Quote:
Originally Posted by dave999z View Post

I unfortunately do not really understand greg's answer to this question. I do not understand why it is better to aim everything for the Rec709 points when you know that one primary and one secondary can't get there. It seems that would result in a distorted triangle. And what's the use of the calculator then (if Rec709 is your goal)?

Many thanks in advance for any assistance anyone can provide.

I think the best way to answer your main question is to ask why you would want to aim at something other than the standard (Rec. 709 or Rec. 601)? If the optimum color accuracy is achieved by matching one of those standards, why would you want to aim for anything else? If you don't have the adjustments necessary to exactly achieve the standard colorimetry, why would you think exactly achieving a wrong colorimetry would be better than getting as close as you can to the right colorimetry? By adjusting for the minimum xyY errors (measured by dE) relative to the standard colorimetry you are effectively minimizing the errors within the 3-D (xyY) color gamut. And that is one of the 4 uses I explained for the calculator.

I'm not sure what you meant when you said "distorted" triangle, but it seems like that is influencing your thinking. Any set of three primaries creates a color gamut triangle in an x,y plane, so I don't see how any choice is "distorted". But the objective to color accuracy is to not only match the x,y coordinates of the target primaries (and complementary colors), but to also match the correct luminance (Y) of the target primaries (and complementary colors). The latter is just as much a contributor to the perceived color accuracy errors as the x,y values. So we need to think in 3-dimensions to measure the color errors and that is what dE does for us. If I have the ability to adjust the Y values of the primaries/complementaries independent of their x,y values, then I should do so with the purpose of minimizing the color accuracy errors relative to the standard, not to simply put the Y values where they would have been on a display with the wrong primaries. Again, what would the point be of exactly achieving a wrong target, rather than getting as close as possible to the right target?

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 04-23-2009, 01:17 PM
AVS Special Member
 
kjgarrison's Avatar
 
Join Date: Feb 2007
Location: Northwest Wisconsin
Posts: 2,243
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Greg, thanks for coming back in and trying yet again.

First off, let me apologize for saying you were annoyed. I just re-read your posts and I think I was projecting. So, again, I apologize.

Another benefit of reading your posts again is that I'm starting to understand what you are saying.

I know: Finally

If I might make a suggestion to you on dealing with us noobs ...

You frequently say things about what we "want". Things about which standard we want to use and things about this or that that we want.

Of course I can't speak for anybody else, but that's entirely too much freedom for me. I need to be TOLD what to do. As to what I, personally, want, it is simple. I want my TV to look as good as it can given that I cannot adjust the primaries xy coordinates. Not at all. I can adjust their Y values, and I can adjust the secondaries xyY usually to whatever it needs to be (I can't push a secondary's xy outside of the triangle formed by the primaries, but I can pull it inside.)

When I use your calculator with inputs for the "custom" primary colors' xy coordinates, I get lines from the primaries to their respective secondaries that intersect nicely and reassuringly at the white point. The secondaries are not, of course, where the Rec709 (the standard I choose) secondaries are located, but the primaries and their secondaries are lined up right with respect to white (D65).

Of course on re-reading your posts above, I now see and understand that you said that is exactly how it is supposed to work. And that this just means that my TV is doing it's color decoding in a standard way, making this more of a test of how my TV works than a tool for me to adjust it's faulty colors.

And I also see that you say it is my option if I want to, to adjust everything to these custom targets. However, you recommend adjusting everything that can be adjusted to the standard of my choice, in my case Rec709. This will result in absence of convergence at the white point, as I'm sure you know.

Anyway, thanks, and again please accept my apology.
kjgarrison is offline  
Old 04-23-2009, 01:41 PM
Senior Member
 
dave999z's Avatar
 
Join Date: Jan 2009
Posts: 216
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 16
Greg, thanks for the thoughts... appreciated.

Quote:
Originally Posted by gregr View Post

Again, what would the point be of exactly achieving a wrong target, rather than getting as close as possible to the right target?

My tv's green cannot get to the 709 xy point. Empirically, I seem to have achieved better skin tones using the calculator and aiming for the "exactly wrong target" xy and luminances, than when I just aim for the 709 spec. (And I think some others have too.)

From what you explain, it's probably the adjustment to the calculated primary and secondary luminances (i.e., the z axis) that helped, not necessarily the adjustment to the calculated xy points for the secondaries.

So, maybe the results would be even better by calibrating all primary and secondary xy points as close to 709 as possible, then running the calculation, then calibrating to the computed target luminances, but ignoring the computed secondary xy points (i.e., leave your secondary xy points as close to 709 as they can get). Maybe?

Otherwise, I don't know why the colors look more "right" (for lack of a better word) when I calibrate to the computed xy values than when I calculate to the 709 xy values.
dave999z is offline  
Old 04-23-2009, 02:55 PM
Senior Member
 
dave999z's Avatar
 
Join Date: Jan 2009
Posts: 216
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 16
Quote:
Originally Posted by kjgarrison View Post

And I also see that you say it is my option if I want to, to adjust everything to these custom targets. However, you recommend adjusting everything that can be adjusted to the standard of my choice, in my case Rec709. This will result in absence of convergence at the white point, as I'm sure you know.

Right, so, which is the better approach, and why?
dave999z is offline  
Old 04-23-2009, 03:25 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by kjgarrison View Post

I want my TV to look as good as it can given that I cannot adjust the primaries xy coordinates. Not at all. I can adjust their Y values, and I can adjust the secondaries xyY usually to whatever it needs to be (I can't push a secondary's xy outside of the triangle formed by the primaries, but I can pull it inside.)

Now I'm the one that is confused. Are you trying to adjust a TV using just its Color and Hue (Tint) controls? The calculator is really aimed at doing a display calibration with some type of a color management system (CMS). It you only have Color and Hue controls you are extremely limited in what you can do.

Quote:
Originally Posted by kjgarrison View Post

When I use your calculator with inputs for the "custom" primary colors' xy coordinates, I get lines from the primaries to their respective secondaries that intersect nicely and reassuringly at the white point. The secondaries are not, of course, where the Rec709 (the standard I choose) secondaries are located, but the primaries and their secondaries are lined up right with respect to white (D65).

Of course on re-reading your posts above, I now see and understand that you said that is exactly how it is supposed to work. And that this just means that my TV is doing it's color decoding in a standard way, making this more of a test of how my TV works than a tool for me to adjust it's faulty colors.

YES! You have it now. If the lines connecting primaries to their corresponding complementary colors ALL intersect at the white reference, then your YCbCr->RGB color decoder is working in a standard manner (it is "orthogonal" in technical terms) and there are no other "odd" non-linearities in the system. This a pre-requisite to achieving totally accurate color using YCbCr (digital) or YPbPr (analog) input signals. (If you input RGB signals, then this should always be true because there is no color decoding.) Once you have the color decoding correct, it is then necessary to have the correct RGB primary x,y values and the correct white point (and a constant gray scale) to achieve totally accurate color. And that requires a CMS or a display with standard primaries.

If you only have Color and Hue controls (no CMS) then the only things you can do to improve color accuracy if you have non-standard primary colors is to "slightly" modify the white reference point to shift the complementary colors closer to the standard, and/or to use the Hue control to simultaneously rotate all of the complementary colors closer to their standard targets. The calculator can help you minimize the errors when making those adjustments. When you are done with that you will probably not like the flesh colors on every piece of video you watch. You would then adjust the Color control to optimize the appearance of the flesh tones, and you might find the Color control would be a little different for different source material (such is the nature of trying to compromise a non-accurate system).

If you have TV with service level adjustments for the color decoder, you could also modify the color decoding (so it's no longer orthogonal - as defined above) which can also help to improve the overall color accuracy in a system with the wrong primary colors. But that is definitely not something anyone should do unless they have great experience in these matters, because A) you are more likely to make things worse than when you started even if it looks better to you on a few things, and B) if you aren't careful keeping exact track of what you did, you may never get back to the factory setup. This is a real recipe for disaster - you were warned!

Greg Rogers
Video Engineer/Product Designer

gregr is offline  
Old 04-23-2009, 03:36 PM - Thread Starter
AVS Special Member
 
gregr's Avatar
 
Join Date: Nov 2001
Posts: 2,129
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Quote:
Originally Posted by dave999z View Post

My tv's green cannot get to the 709 xy point. Empirically, I seem to have achieved better skin tones using the calculator and aiming for the "exactly wrong target" xy and luminances, than when I just aim for the 709 spec. (And I think some others have too.)

So do I understand that you have a CMS to make these x,y and Y adjustments to the primaries and complementary colors but it doesn't have enough range to adjust the green correctly?

Quote:
Originally Posted by dave999z View Post

From what you explain, it's probably the adjustment to the calculated primary and secondary luminances (i.e., the z axis) that helped, not necessarily the adjustment to the calculated xy points for the secondaries.

So, maybe the results would be even better by calibrating all primary and secondary xy points as close to 709 as possible, then running the calculation, then calibrating to the computed target luminances, but ignoring the computed secondary xy points (i.e., leave your secondary xy points as close to 709 as they can get). Maybe?

Otherwise, I don't know why the colors look more "right" (for lack of a better word) when I calibrate to the computed xy values than when I calculate to the 709 xy values.

Because what people prefer for color can be very subjective and what they find most pleasing is often based on what they have previously watched, which may not have been accurate. Or because what you think was accurate actually wasn't.

Greg Rogers
Video Engineer/Product Designer

gregr 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