Ok I have read numerous posts/threads on calibration and Ive got my head around Greyscale / Gamma curves / Colour Gamut and using an eye one/HCFR I think I have my JVC RS1 set up pretty well. But I just cant get my head around the difference between Luminance as understood when calibrating Greyscale/Gamma and the Luminance/brightness of a primary colour as defined as the 'Y' value.
Surely the luminance 'Y' of a colour effects the grayscale and visa versa, but I get the impression from many posts that they are independant.:confused: I know the 'Y' value is managed by the colour decoder of the device and that the secondary colours are influenced by brightness, but from here on in....?
I am also confused about the logarithmic version of the Gamma curve on HCFR.
Captain Novice here would be grateful for any replies that makes these easier for me to understand.

Color is a 3-dimensional property including saturation, hue and lightness. RGBCYM are colors just as white is a color. With respect to the lightness component, the only difference is that RGBCYM are each defined by a specific lightness value. White is not. That's why it is called a gray scale. Each level in the gray scale should have the same hue/saturation value (0.3127, 0.329), but different values of lightness. For a specified gamma, you could define a specific lightness standard for each of the levels on the gray scale. Thus, whereas one could say that a particular representation of red contains a lightness error, the same cannot be said about white without first specifying the % stimulation (10-100 IRE) and the desired gamma.

Ok I have read numerous posts/threads on calibration and Ive got my head around Greyscale / Gamma curves / Colour Gamut and using an eye one/HCFR I think I have my JVC RS1 set up pretty well. But I just cant get my head around the difference between Luminance as understood when calibrating Greyscale/Gamma and the Luminance/brightness of a primary colour as defined as the 'Y' value.
Surely the luminance 'Y' of a colour effects the grayscale and visa versa, but I get the impression from many posts that they are independant.:confused: I know the 'Y' value is managed by the colour decoder of the device and that the secondary colours are influenced by brightness, but from here on in....?
I am also confused about the logarithmic version of the Gamma curve on HCFR.
Captain Novice here would be grateful for any replies that makes these easier for me to understand.

Thanks Tom . I understand the majority of your reply, thanks. So are you saying that if you increased the 'Y' value on say the green primary, which I assume you are referring to as the 'lightness error' that it will have no effect on the green representation on a greyscale graph ?
Now as the Y value of the primaries should be, from memory, green 71% red 24% red 8%of white I assume these %ages should be the same for all stimulus values. But what if as in the case of the RS1 the green is over 100% of white can this be compensated for by a different greyscale ? or have I got it completely wrong

Thanks Tom . I understand the majority of your reply, thanks. So are you saying that if you increased the 'Y' value on say the green primary, which I assume you are referring to as the 'lightness error' that it will have no effect on the green representation on a greyscale graph?Correct.


Now as the Y value of the primaries should be, from memory, green 71% red 24% red 8%of white I assume these %ages should be the same for all stimulus values. But what if as in the case of the RS1 the green is over 100% of white can this be compensated for by a different greyscale ? or have I got it completely wrongFirst, it is 21% for red. Second, the RS1 green error is NOT in Y it, but in saturation, which is measured by a x,y coordinate.

If you are getting a Y error for the RS1, this is a known color space matching problem, which can be resolved by running RGB instead of YCbCr.

Correct.


Tom, I must disagree with you on this...although the test input/source of any IRE/stimulus level is pure luminance absent any chroma it is not the case that changing the luminance level of a primary will have no impact on the grayscale calibration as the primary -- let's continue to use green -- that being added to get the desired white point of (0.3127, 0.3290) is different thus the mix of the other primaries (i.e. red and green) will be different...

Your comments/thoughts?

it is not the case that changing the luminance level of a primary will have no impact on the grayscale calibration as the primary -- let's continue to use green -- that being added to get the desired white point of (0.3127, 0.3290) is different thus the mix of the other primaries (i.e. red and green) will be different...

Joelc,

The defiinition implies to me that luminance is not a factor in defining the white point. Only x and y define it. If you were, for instance, to increase the red primary luminance to excess, then you would have "red push" as I understand it. The gray scale could be correct and yet have a reddish cast due to the excess luminance. But I could be wrong.

Mike

Correct.


First, it is 21% for red. Second, the RS1 green error is NOT in Y it, but in saturation, which is measured by a x,y coordinate.

If you are getting a Y error for the RS1, this is a known color space matching problem, which can be resolved by running RGB instead of YCbCr.

Joelc,

The defiinition implies to me that luminance is not a factor in defining the white point. Only x and y define it. If you were, for instance, to increase the red primary luminance to excess, then you would have "red push" as I understand it. The gray scale could be correct and yet have a reddish cast due to the excess luminance. But I could be wrong.

Mike


To me there are three things that are getting crossed here:

-- Luminance is not a factor in defining grayscale OTHER THAN THE FACT that grayscale is calibrated/measure from 0% IRE/stimulus to 100% IRE/stimulus generally in 10% IRE/stimulus increements...but, I do agree, that grayscle test patterns SHOULD contain no colour (i.e. Y > 0, Cb=0, Cr=0).

-- Calibrating grayscle involves "changes the balance" of R, G and B such that the "mixture" provides the defined reference point for the given colour gamut (i.e. (x,y) = (0.3127, 0.329) for SMPTC-E and REC 709)...now, let's suppose that the "starting" setting for R, G and B provide perfect grayscale...is it not the case that changing -- say the luminance of G -- will change the result of the mixing R, G and B because the G that one is adding has changed and thus result in the need to recalibrate the grayscale?

-- And then there is gamma, think of gamma as the grayscale calibration of luminance...

HTH

As I pointed out in the CMS calibration thread, although color decoding, gray scale, and gamut are completely separate adjustments, there MAY be some interactive effect. However, the effect will be small and unpredictable.

So, in general color decoding settings (primary lightness) are independent of the gray scale (xy of white), but depending upon the display adjusting one may affect the other at the margins.

is it not the case that changing -- say the luminance of G -- will change the result of the mixing R, G and B because the G that one is adding has changed and thus result in the need to recalibrate the grayscale?Probably not, because what you are adding is a lightness component, which is not part of grayscale measurement.

Probably not, because what you are adding is a lightness component, which is not part of grayscale measurement.

Fair enough...but it will then the gamma...

Fair enough...but it will then the gamma...It might. You'd have to measure it directly on a case-by-case basis to say for sure.

Tom. I found this quote of yours which perhaps explains my missunderstanding better;

- Perhaps I was most surprised by the RS1's color performance. Red is quite oversaturated and green is extremely oversaturated, as has been reported many times. However, what shocked me was the color decoder. Green is exaggerated by 38%. Red and blue were about 10% lower than ideal levels. The is the most extreme color luminance error I've ever encountered on any display of any type. Thus, without a dedicated color decoder adjustment or a full CMS, this was impossible to adjust out. Getting green anywhere near normal resulted in severe loss of brightness for red and blue.-

I ran some measurements using HCFR with my own RS1. Once recalibrated for grayscale and Gamma and with the RS1s colour control on 0, I observed the following;

Primary colours measured at ;
Red 31% of white, Green 107% of white, and Blue 8% of white. But with the RS1s colour control set to -30 you get ;
Red 21%, Green 96 %, and Blue 3.5% of white. So the Red is bang on but with the green over and blue under and with the primaries fairly close to their correct positions on the CIE chart.

Now having read that the new Radience can reduce saturation whilst leaving colour brightness alone, and hence far too high on red and green on the RS1, surely the overbrightness washes out the colors. So is the Radience a real solution to the RS1s problems ?

Thanks for your time and training !

It would help if I knew where you were quoting this from.

The error reported in this quote and seen by you almost certainly refers to a color space mismatch problem that is well known with the RS1. Run RGB only and the lightness problem should go away.

Yes, the Radiance seems like a good solution. I also good pretty good results with the Crystalio.

Tom. I found this quote of yours which perhaps explains my missunderstanding better;

- Perhaps I was most surprised by the RS1's color performance. Red is quite oversaturated and green is extremely oversaturated, as has been reported many times. However, what shocked me was the color decoder. Green is exaggerated by 38%. Red and blue were about 10% lower than ideal levels. The is the most extreme color luminance error I've ever encountered on any display of any type. Thus, without a dedicated color decoder adjustment or a full CMS, this was impossible to adjust out. Getting green anywhere near normal resulted in severe loss of brightness for red and blue.-

I ran some measurements using HCFR with my own RS1. Once recalibrated for grayscale and Gamma and with the RS1s colour control on 0, I observed the following;

Primary colours measured at ;
Red 31% of white, Green 107% of white, and Blue 8% of white. But with the RS1s colour control set to -30 you get ;
Red 21%, Green 96 %, and Blue 3.5% of white. So the Red is bang on but with the green over and blue under and with the primaries fairly close to their correct positions on the CIE chart.

Now having read that the new Radience can reduce saturation whilst leaving colour brightness alone, and hence far too high on red and green on the RS1, surely the overbrightness washes out the colors. So is the Radience a real solution to the RS1s problems ?

Thanks for your time and training !

Now having read that the new Radience can reduce saturation whilst leaving colour brightness alone, and hence far too high on red and green on the RS1, surely the overbrightness washes out the colors. So is the Radience a real solution to the RS1s problems ?

008,

My luminance errors were quite low using an earlier firmware version of the Radiance: http://www.avsforum.com/avs-vb/showthread.php?p=12622116#post12622116 . However, recent changes in the firmware resulted in a perfect gray scale calibration, but with huge luminance errors. I reverted back to the earlier firmware while Lumagen sorts out the issue.

Mike

It would help if I knew where you were quoting this from.

-The error reported in this quote and seen by you almost certainly refers to a color space mismatch problem that is well known with the RS1. Run RGB only and the lightness problem should go away.

Yes, the Radiance seems like a good solution. I also good pretty good results with the Crystalio.

Sorry Tom, your quote was a post on Harry Brants RS1 back in May.

My numbers as quoted above where run in RGB from Crystalio 2 test patterns in to the JVC. Perhaps you could confirm I am measuring correctly; I am doing a 20 point greyscale run and also measuring the primaries using HCFR, now, with numeric values in xyY I am dividing each colour Y vaue by the Y value of 75% white as taken from the greyscale run. Could you confirm this is the correct proceedure thanks.

You say you are getting pretty good results with the Crystalio. Can you tell me roughly how you have set yours up with regard to colour as I am not getting good results with mine (probably not using in correctly) :o

Cheers

Sorry Tom, your quote was a post on Harry Brants RS1 back in May.And if you read on in that thread you'll find that it was a color space mismatch problem that we resolved.

http://www.avsforum.com/avs-vb/showpost.php?p=10487206&postcount=85

You say you are getting pretty good results with the Crystalio. Can you tell me roughly how you have set yours up with regard to colour as I am not getting good results with mine (probably not using in correctly)First, make sure that the RS1 is set up to receive RGB, not YCbCr.

Second, are you sure that the Crystalio's test patterns are 75% patterns? Use external patterns to be sure.

Third, color calibration on the Crystalio is something of a pain, but it does work. Begin by lowering the saturation of the primaries by adding complementary colors to the oversaturated primary (R and B to G, for example). Now the xy is good, but the lightness is too low. Now use the primary saturation control to fix this. This will raise the lightness, but it ALSO increases saturation, so you will lose some of what you accomplished in the first step. However, you can get it MUCH better than it was to begin with.

And if you read on in that thread you'll find that it was a color space mismatch problem that we resolved.
http://www.avsforum.com/avs-vb/showpost.php?p=10487206&postcount=85
First, make sure that the RS1 is set up to receive RGB, not YCbCr.
Second, are you sure that the Crystalio's test patterns are 75% patterns? Use external patterns to be sure.
Third, color calibration on the Crystalio is something of a pain, but it does work. Begin by lowering the saturation of the primaries by adding complementary colors to the oversaturated primary (R and B to G, for example). Now the xy is good, but the lightness is too low. Now use the primary saturation control to fix this. This will raise the lightness, but it ALSO increases saturation, so you will lose some of what you accomplished in the first step. However, you can get it MUCH better than it was to begin with.


Sorry Tom I did not see that later post.

I definately am using RGB on the RS1

Yes I have compared the Crystalios patterns with DVE and they are pretty much identical. DVE states that the colour patterns are 75%

Thanks for the tip Tom. I will give that a go.

""My numbers as quoted above where run in RGB from Crystalio 2 test patterns in to the JVC. Perhaps you could confirm I am measuring correctly; I am doing a 20 point greyscale run and also measuring the primaries using HCFR, now, with numeric values in xyY I am dividing each colour Y vaue by the Y value of 75% white as taken from the greyscale run. Could you confirm this is the correct proceedure thanks"".

Tom, sorry to be a calibration newbie pain but can you confirm that I am using the correct proceedure for calculating primary colour lumnance.

Thanks Tom.

You have to make sure that you are BOTH outputting RGB from the Crystalio AND that the JVC is setup to receive RGB. It should be fine.

Sorry Tom I did not see that later post.

I definately am using RGB on the RS1

Yes I have compared the Crystalios patterns with DVE and they are pretty much identical. DVE states that the colour patterns are 75%

Thanks for the tip Tom. I will give that a go.

""My numbers as quoted above where run in RGB from Crystalio 2 test patterns in to the JVC. Perhaps you could confirm I am measuring correctly; I am doing a 20 point greyscale run and also measuring the primaries using HCFR, now, with numeric values in xyY I am dividing each colour Y vaue by the Y value of 75% white as taken from the greyscale run. Could you confirm this is the correct proceedure thanks"".

Tom, sorry to be a calibration newbie pain but can you confirm that I am using the correct proceedure for calculating primary colour lumnance.

Thanks Tom.

Third, color calibration on the Crystalio is something of a pain, but it does work. Begin by lowering the saturation of the primaries by adding complementary colors to the oversaturated primary (R and B to G, for example). Now the xy is good, but the lightness is too low. Now use the primary saturation control to fix this. This will raise the lightness, but it ALSO increases saturation, so you will lose some of what you accomplished in the first step. However, you can get it MUCH better than it was to begin with.

Just to clarify Tom, the Crystalio colour calibration suite comprises of three sliders for each primary colour so to adjust for green you have three sliders for red green and blue.
There is also the primary SATURATION control which affects all colours and in addition three independant SATURATION controls for red green and blue.

When you say add complementary colours to the oversaturated green on the RS1 are you refering to the colour controls or the independant red green and blue saturation controls as referred to above ?

Once again many thanks.

Hi Tom
Would be grateful for a reply on this when you get chance. Once again many thanks.

To desaturate green use the sliders for green, adding both blue and red. Then you'll have to increase the green saturation to restore some of the lost lightness from the previous step.