or Connect
AVS › AVS Forum › Display Devices › Display Calibration › Display Calibration Calculator
New Posts  All Forums:Forum Nav:

Display Calibration Calculator

post #1 of 136
Thread Starter 
I've posted PC and Mac OS X versions of a Display Calibration Calculator tool on the AccuPel website. You can download them at the bottom of the page that has the HDG-4000 manuals.

The Display Calibration Calculator includes the features of the previous Luminance Calculator & Color Gamut Calculator plus adds Grayscale dE and gamma calculations.

Features:

1. Primary and Complementary CIE x,y,Y values for Rec. 709, Rec. 601, and PAL standards, or any user-entered WRGB CIE x,y values.

2. dE values of measured WRGBYCM x,y,Y values vs. the above target values.

Note: The measured Y value of Gray will not affect the dE value of Gray because it is used as a reference to calculate other dE values. i.e. it is assumed to be the desired Y value for Gray. But it will affect the dE values of the other colors. ---- You can enter any value of Y for Gray in the Target section to calculate all of the other color's Y target values.

3. Grayscale dE calculation
4. Grayscale Gamma curve calculation

The latter also provides Y (luminance) target values for a selected gamma to adjust a processor that provides custom gamma curves. You can optionally enter a 0% (black field) luminance value or Full-field (on-off) Contrast Ratio to compute the display's signal gamma curve (i.e. remove the projector's black level from the gamma calculations).

Enjoy!
post #2 of 136
Thread Starter 
Updated (to version 1.21) to add a custom icon supplied by WolfyA (thanks). That is the only change.
post #3 of 136
You are very welcome Greg.

Thanks for your added functionality, keeps getting better and better.

Any chance you could add a "Always on top" feature?

For example add an extra menu item, e.g. "View" after "File" and "Edit", and then add "Always on top" to that menu so that, when selected, the calculator always stays in front regardless of what window is in focus. Would help a lot when one's using copy and paste to enter values.

Cheers.
post #4 of 136
This is a great little tool! I've used the Luminance Calculator ever since I learned about it. Question- where can I read/learn about color calibrating? Specifically, as to why secondary colors should not be calibrated to standard if primary colors are off. I find it interesting.

Thanks!
post #5 of 136
HDholic
Quote:


This is a great little tool! I've used the Luminance Calculator ever since I learned about it. Question- where can I read/learn about color calibrating? Specifically, as to why secondary colors should not be calibrated to standard if primary colors are off. I find it interesting.

The secondary values are calculated from the primaries so if the measured primaries are incorrect then the secondaries can not be valid either. You must calculate the secondaries from the measured primaries which will allow you to adjust the chroma decoder accurately. If your display has CMS (Color Management Software) then you should be able to correct the primaries correctly however secondary correction in color management does not always work well as it is simply adjusting two primaries.

We provide a tool in our CA-6X software which performs the calculation for the measured primaries to correctly display in the CIE graph where the secondaries should be located. We have recently updated our software to include a calculator which performs all of the functions in Greg Rogers calculator automatically (when connected to either an Accupel HDG-4000 or Sencore 4XX series generator) which logs measurements from the color analyzer for primaries, secondaries, and White and then displays the target luminance/illuminance values for the BR-709, BT-601, and EBU/PAL colorspace. The data is plotted and shows the correct "Y" value on the graph at 100 points after the user selects which colorspace they wish to have displayed in the graph. The calculator also allows the user to enter the desired target gamma value which is factored into the calculation. Measurements may be entered manually if neither generator is connected to the software however this could be fairly time consuming to perform!

Performing the measurement in ten ire increments for all seven colors takes approximately 4-5 minutes depending on the probe being used for the test.
post #6 of 136
I have been using this calculator, and I have all colors with dE < 3. But for some reason the holy grail of TV, human skin, still doesn't look right.

Samsung LNT 4671 LCD. The CIE chart color gamut for this TV is reasonably close for blue and red, but green is off to the left. There is a setting in these Sammys in "My Color Control" called "White". Moving this white slider moves white along the grayscale line. A new White setting requires re-doing of the grayscale.

But I notice that this White slider seems to shift the whole white point <-> CIE relationship.

Since green is off to the left, the line from green through the white point defines magenta on the line between blue and red. Changing "White" on that slider in My Color Control is like a small step color temp control, going between red and blue, requiring re-do of the grayscale. And, since it moves almost parallel with the blue-red line, the secondary that is most effected is magenta. Only one of these White slider settings results in the "right" magenta. And that setting, along with grayscale causes all three lines that define secondaries to cross at the white point.

But the Accupel calculator does more than show the correct xy coordinates for each secondary for any given set of primaries. It calls for different Y values than the usual for the Rec709 gamut. In my case, since green is where it is, the calculator calls for a higher red Y, closer to 25% of white Y instead of 21%.

So, based on my rudimentary understanding, if we have TV's without CMS and the ability to put the primary colors at spec (and many of us do), the Accupel calculator calls for ("calls for" meaning gives lower dE values when) the white point D65 (x=.312 and y=.329), defines coordinates of non-standard secondaries based on the relationship of D65 and the non-standard primaries, and calls for a non-standard luminance (Y) of all colors.

I suppose there could be innumerable ways to adjust any 3 non-Rec709 primary colors, including innumerable different white points, given the ability to adjust luminance of the colors. But all the software is set up for D65 white point, so why not just go ahead and stick with that.

And it appears that Samsung engineers agree with the method this Accupel Calculator uses. Using the luminance ratios (like red Y = 21% white Y) and primary/secondary color locations that apply to Rec709 perfect color gamut, my TV has major problems and BIG dE's all over the place. But using the unique color gamut it comes with, it automatically puts out luminance values close to perfectly in accordance with the Accupel calculator.

Well done from the numbers/measurements standpoint. Maybe I just need a little more time to get used to it. Some skin looks perfect. Other ppls skin looks like The House Of Wax
post #7 of 136
Quote:
Originally Posted by kjgarrison View Post

But I notice that this White slider seems to shift the whole white point <-> CIE relationship.

kj,

Do you mean that if you alter your grayscale, that your secondaries move? This is true since the position of the secondaries depends on the white point. I certainly found on a couple older CRTs that had very bluish white points that after grayscale calibration, the secondaries had moved quite noticeably - which was expected.

Quote:


Only one of these White slider settings results in the "right" magenta. And that setting, along with grayscale causes all three lines that define secondaries to cross at the white point.

I'm a bit confused. Are you saying that you altered your white point (i.e. it's not D65) in order to get magenta to be on target? Or are you saying that you were able to get a better grayscale tracking by using the White MyColorControl in addition to the White Balance RGB controls and this also happened to get magenta on target at the same time? I have not touched the White setting in MyColorControl thinking that it was simply shifting the entire grayscale thus starting from Warm 1 or Warm 2 and using the White Balance RGB controls was sufficient. I'm just curious if you found that you were able to get closer to D65 by first using the White MyColorControl to shift things a bit and then using the White Balance controls to fine tune from there.

Quote:


I suppose there could be innumerable ways to adjust any 3 non-Rec709 primary colors, including innumerable different white points, given the ability to adjust luminance of the colors. But all the software is set up for D65 white point, so why not just go ahead and stick with that.

All brightness targets and secondaries are derived from a set of primaries and a reference white point. This is true whether or not you're a standards body defining HD709 or you're defining own custom color gamut. So, it is up to you to select the white point and the primaries in order to derive everything else. When using a gamut calculator, I have been simply entering in D65 as my reference white point since that it what I am shooting for when calibrating grayscale.

Quote:


Well done from the numbers/measurements standpoint. Maybe I just need a little more time to get used to it. Some skin looks perfect. Other ppls skin looks like The House Of Wax

Without a known reference monitor to compare the viewing material, it's hard to say how it is supposed to look but if your colors are off from either SMPTE-C or HD709 then it will look different on your display from what it most likely looked like on monitors used to master the material. In the end, this all speaks to the need for manufacturers to include decent CMS controls and/or to engineer their displays to be more adherent to accepted standards.

hope this helps,


--tom
post #8 of 136
from extensive experience....

If the CMS controls do not allow for manipulation of x y and Y color points then you may just be better off leaving them in the default position and doing the white balance and color and tint settings (based on skin tones).

YMMV

Gregg
post #9 of 136
Quote:
Originally Posted by HDholic View Post

This is a great little tool! I've used the Luminance Calculator ever since I learned about it. Question- where can I read/learn about color calibrating? Specifically, as to why secondary colors should not be calibrated to standard if primary colors are off. I find it interesting.

You could start here.

The calculation for the secondaries is essentially geometrical. The correct point for a secondary is the intersection of the line between the contributing primaries and a line drawn between the white point and the complementary primary. This can be expressed mathematically.
post #10 of 136
I'm hoping this is a simple question. Because my primaries are off, I need to use the calculator to create targets for my secondaries. I usually use my measured values for white which are usually close to the ideal of 0.3127, 0.3290. For example when I measured it last they were 0.314, 0.331 (very close but not exactly ideal). Is it correct to use the measured values for white when calculating the new target secondaries or should I use the ideal. It seems to me that the measured white point should be used because this is a calculation based on a measured primary, straight line (through the measured white) to a new target secondary. Thanks for your input and a great big huge thanks to Greg for creating such a great tool!
post #11 of 136
you need to use the measured white point with the measured primaries.
Please keep this is mind:
if you dont have x y Y adjustments for your secondaries, it is doubtful you will have a visible improvement and may actually make the image look less correct (even though your secondary targets may measure closer to the target points).
post #12 of 136
Thanks, Gregg. I don't have xyY adjustments for the secondaries, just tint. I found the most accurate is leaving it at the default of zero but adjusting the color based on Y.
post #13 of 136
Quote:
Originally Posted by Gregg Loewen View Post

you need to use the measured white point with the measured primaries.

Gregg, what if your white point varies across the scale yet the primaries remain fixed? Do you still just use the white point measured (along with RGBYCM) with the 75% or 100% patterns?


--tom
post #14 of 136
Quote:
Originally Posted by thomasl View Post

kj,

Do you mean that if you alter your grayscale, that your secondaries move? This is true since the position of the secondaries depends on the white point. I certainly found on a couple older CRTs that had very bluish white points that after grayscale calibration, the secondaries had moved quite noticeably - which was expected.



I'm a bit confused. Are you saying that you altered your white point (i.e. it's not D65) in order to get magenta to be on target? Or are you saying that you were able to get a better grayscale tracking by using the White MyColorControl in addition to the White Balance RGB controls and this also happened to get magenta on target at the same time? I have not touched the White setting in MyColorControl thinking that it was simply shifting the entire grayscale thus starting from Warm 1 or Warm 2 and using the White Balance RGB controls was sufficient. I'm just curious if you found that you were able to get closer to D65 by first using the White MyColorControl to shift things a bit and then using the White Balance controls to fine tune from there.

--tom

Tom, first let me apologize for missing your response to my questions. I do subscribe to this thread, but I must have missed your reply.

The secondaries (primarily magenta) end up in different places with different settings of the white slider in MCC after recalibration of the grayscale.

Here is what that white slider in MCC does, or at least what it seems to do. While you move it, the overall appearance gets more blue or more white. If you do a continuous reading of white while moving the slider it tracks along the grayscale, and it looks bluer or redder depending on which way it is moved.

Changing this setting seems like smaller adjustments than are possible with the color temp settings (standard, warm1, warm2, etc.). After changing this slider, the grayscale has to be set right (back to D65). BUT the position of the secondaries, especially magenta is different. It's sort of like the whole CIE triangle is shifted mostly left or right (since the grayscale tracking is more or less parallel to the blue-red line) by this white slider.

Bottom line the fact that green primary is so far off effects magenta much more than the other secondaries. Moving this slider to different points, and then re-calibrating the grayscale results in magenta ending up to the left, to the right, or close to right on, it's "proper" CIE spot.

Quote:
Originally Posted by thomasl View Post

kj,

All brightness targets and secondaries are derived from a set of primaries and a reference white point. This is true whether or not you're a standards body defining HD709 or you're defining own custom color gamut. So, it is up to you to select the white point and the primaries in order to derive everything else. When using a gamut calculator, I have been simply entering in D65 as my reference white point since that it what I am shooting for when calibrating grayscale.



Without a known reference monitor to compare the viewing material, it's hard to say how it is supposed to look but if your colors are off from either SMPTE-C or HD709 then it will look different on your display from what it most likely looked like on monitors used to master the material. In the end, this all speaks to the need for manufacturers to include decent CMS controls and/or to engineer their displays to be more adherent to accepted standards.

hope this helps,


--tom

I always use D65 (or as close to it as I can get )
post #15 of 136
Quote:
Originally Posted by kjgarrison View Post

The secondaries (primarily magenta) end up in different places with different settings of the white slider in MCC after recalibration of the grayscale.

kj,

thanks for the explanation. I had done some measures/experiments with the white color control on my Samsung LCD. It seems to do what yours does as well. It acts as a way to more finely tune the color temperature setting - either more blue or red. What I didn't do is see what happened to the secondaries if I put the white control in a setting other than 15 and re-calibrated grayscale. Am I correct in thinking that if you moved the white control so it made things bluer that it moved your magenta to the left after you re-calibrated with the White Balance controls?

My green primary is actually not that far off - just under-saturated but still pretty close to the line to the white point so my magenta can be made fairly accurate.

cheers,


--tom
post #16 of 136
Quote:
Originally Posted by thomasl View Post

kj,

thanks for the explanation. I had done some measures/experiments with the white color control on my Samsung LCD. It seems to do what yours does as well. It acts as a way to more finely tune the color temperature setting - either more blue or red. What I didn't do is see what happened to the secondaries if I put the white control in a setting other than 15 and re-calibrated grayscale. Am I correct in thinking that if you moved the white control so it made things bluer that it moved your magenta to the left after you re-calibrated with the White Balance controls?

My green primary is actually not that far off - just under-saturated but still pretty close to the line to the white point so my magenta can be made fairly accurate.

cheers,


--tom

I actually don't remember which way magenta moves, but I think it is as you described with lower white slider setting resulting in magenta movng to the left. As I recall, lower white slider settings makes white look more red. Maybe it's backwards to that. I never put it all together in my mind. You can do a continous reading on white as you change the white slider. You will see that the "white" spot moves along the grayscale tracking and that the overall color of white gets bluer or redder. But then when you recal the grayscale and get white back to D65, magenta will have moved. Like I said I don't remember exactly, but I believe lower white settins results in magenta moving to the left.

As to whether this setting is fine tuning color temp, I'm not so sure that's all it does. As far as I know, different color temps with grayscale adjustments don't result in different locations of secondaries (particularly magenta, since color temps involve red and blue). I never really checked to see if different temps result in different magenta plotting, though.
post #17 of 136
Quote:
Originally Posted by kjgarrison View Post

As to whether this setting is fine tuning color temp, I'm not so sure that's all it does. As far as I know, different color temps with grayscale adjustments don't result in different locations of secondaries (particularly magenta, since color temps involve red and blue). I never really checked to see if different temps result in different magenta plotting, though.

Kj, thanks for the information. Sorry, I just meant that it was a way for the user to fine tune in addition to the color temperature control. The control itself, I'm just guessing, sounds like it is shifting the displays underlying white point - from there all the layers are applied - color temperature offsets and end user white balance offsets. This may explain why the secondaries shift even after you've re-calibrated to D65. Without knowing how it was engineered, it's impossible to say exactly. My (perhaps incorrect) thinking/guessing was as the white control moved things toward blue, magenta would move left toward blue as well since if you draw a line through the new bluer white point from the green primary, you should end up left of where you were.

cheers,


--tom
post #18 of 136
Quote:
Originally Posted by thomasl View Post

Kj, thanks for the information. Sorry, I just meant that it was a way for the user to fine tune in addition to the color temperature control. The control itself, I'm just guessing, sounds like it is shifting the displays underlying white point - from there all the layers are applied - color temperature offsets and end user white balance offsets. This may explain why the secondaries shift even after you've re-calibrated to D65. Without knowing how it was engineered, it's impossible to say exactly. My (perhaps incorrect) thinking/guessing was as the white control moved things toward blue, magenta would move left toward blue as well since if you draw a line through the new bluer white point from the green primary, you should end up left of where you were.

cheers,


--tom

Well, that is how it behaves. As to how it does it, I am not an engineer and have no idea. But it was a useful thing and something that I wish the 750 has. My 71 is bye-bye.
post #19 of 136
Quote:
Originally Posted by kjgarrison View Post

Well, that is how it behaves. As to how it does it, I am not an engineer and have no idea. But it was a useful thing and something that I wish the 750 has. My 71 is bye-bye.

Don't the new Samsung plasma and LCD models from this year have a CMS via the Custom color space option? Are you upgrading or did the 71 have other problems?

cheers,


--tom
post #20 of 136
Quote:
Originally Posted by thomasl View Post

Don't the new Samsung plasma and LCD models from this year have a CMS via the Custom color space option? Are you upgrading or did the 71 have other problems?

Tom, the plasmas definitely do, but it appears that the LCDs do not.
post #21 of 136
Quote:
Originally Posted by TomHuffman View Post

Tom, the plasmas definitely do, but it appears that the LCDs do not.

Thanks Tom. That is disappointing news since I swear I had gone to the Samsung web site awhile back and looked over the user manuals for their new 2008 LCDs and it included the Custom color space feature. It wouldn't be the first time in my experience with technology that the manual didn't match what I saw in real life so I'm not surprised. Perhaps they plan to put the feature in with later releases via new firmware.

cheers,


--tom
post #22 of 136
Quote:
Originally Posted by thomasl View Post

Don't the new Samsung plasma and LCD models from this year have a CMS via the Custom color space option? Are you upgrading or did the 71 have other problems?

cheers,


--tom

The 71 had other problems and Samsung offered to exchange for a LN46A750.

The 750 (and the 650) have this Custom Color space option. I don't know if it qualifies as a full CMS, but it has a lot of capabilities.

You can adjust luminance up or down of all (RGBYCM) colors.

You can adjust xy coordinates of all colors also, but within some limits. Each color RGBYCM has 3 sliders for RGB that can be adjusted.

Couple of examples: Like every color, Red has adjustments for Red, Green and Blue. The default is Red=50, Green=0, Blue=0. If Red is oversaturated and needs to be moved towards blue or green, then you can do that by adding some blue or green until you get it where you want it. If this makes the luminance of Red too high, you can reduce Red some, and maybe take away a little of the blue or green. You can get Red where you want it and at the right luminance. But if Red is undersaturated and needs to be moved away from Blue or Green, you can't since they are already at 0 and can't go lower.

In the Samsung Green is always too far to the left. A lot of Red can be added to green and the Green spot moves towards Red, getting closer to it's CIE spot, although slightly undersaturated.

Cyan by default is Red=0, Green=50, Blue=50. So Cyan can be moved back and forth between Blue and Green, and it can be pulled over towards Red if needed. If the Green primary is left off to the left of it's proper CIE spot, Cyan cannot be moved left to the line between Blue and Green, since Red starts off a 0. Again, by adjusting in unison (more or less) whatever colors are required to get Cyan to the desired spot, Cyan luminance can be adjusted quite precisely.

Since the (green and blue) primaries cannot be put perfectly on their CIE spots, I use the Accupel calculator to see what luminance (Y) values the primaries and secondaries should have as well as where they all should be attempted to be placed.

Is this the correct use of the Accupel calculator, btw? I ask this because another poster is telling me that this isn't how the secondary colors should be placed. Basically he says that, for example, Magenta should be placed on it's CIE spot, while the Accupel calculator assigns a lower dE to placement of Magenta not on it's "perfect" CIE spot but opposite the white spot (D65) on the line between Blue and Red. He agrees that this is the 'definition' of Magenta, the secondary color of Green, but he says that the correct color of Magenta will only look right if it is on the perfect CIE location and that a more perfect recreation of the original material can only come from using the CIE spec coordinates for the secondary colors, regardless of where the primary colors' coordinates are.

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.
post #23 of 136
Quote:
Originally Posted by kjgarrison View Post

The 750 (and the 650) have this Custom Color space option. I don't know if it qualifies as a full CMS, but it has a lot of capabilities.

Kj, thanks for the info. I just checked the 5 Series LCD user manuals and those also have it documented as an Advanced picture option.

Quote:


Couple of examples: Like every color, Red has adjustments for Red, Green and Blue. The default is Red=50, Green=0, Blue=0. If Red is oversaturated and needs to be moved towards blue or green, then you can do that by adding some blue or green until you get it where you want it. If this makes the luminance of Red too high, you can reduce Red some, and maybe take away a little of the blue or green. You can get Red where you want it and at the right luminance. But if Red is undersaturated and needs to be moved away from Blue or Green, you can't since they are already at 0 and can't go lower.

So, it sounds like the Red slider component of Red only affects luminance and doesn't move red further/closer to the white point (i.e. affects saturation). And I would assume that this is true of the other two primaries as well. So, if red is oversaturated yet not leaning at all toward blue or green (just too far from D65), is it not possible to move it at all in toward D65? It makes sense that if red is undersaturated you can't move it out. If the red being displayed is the native red of the display then there is no way to make red "more" red. i.e. there is no way to move it out away from D65 nor a way to take away blue or green (affect the hue of red) since there is no blue or green to begin with in the native red.

Quote:


Since the (green and blue) primaries cannot be put perfectly on their CIE spots, I use the Accupel calculator to see what luminance (Y) values the primaries and secondaries should have as well as where they all should be attempted to be placed.

As far as I know, that is the correct use the color gamut calculator. A set of primaries and a white point define everything else - your secondaries and all your luminance values. So, for example, for magenta, you can draw a line from your green primary through your white point (D65) and where that line strikes the line between red and blue is where magenta is placed. This can all be represented mathematically and programmed in software which is what the Accupel color gamut calculator is doing.

It is another question whether or not you want a "perfect" custom color space if that color space is far off from the HD709 reference targets. I've been thinking about this question as well and it is true that if they are far off, your secondary colors, while technically accurate for your custom color space will not look right since they don't match the color space with which the material was mastered. That is true for any display that does not match the reference target perfectly. Whether it is noticeable to you is still another question. A still more interesting question is how the display uses these points to physically create the rest of the colors that lie inside the gamut. So, for example, if your primaries are well off the HD709 but your secondaries match HD709, does this create any oddities in the display of the rest of the gamut - are properly positioned secondaries more important than properly positioned primaries when it comes to re-creating the rest of the colors. These may be questions that only the engineers and software developers who work on the displays can answer. But if anyone with more knowledge and experience wants to chime in, feel free.

For me personally, I've just tried to adjust the overall color luminance values to match my custom gamut values for each display via my lone Color control. Since I also only have a Tint control to affect the secondaries, I've actually just tried to get them to a position where the total dE is minimized for all 3. My primaries aren't that far off from the HD709 reference so there isn't a big difference between HD709 and my "custom" color gamut.

cheers,


--tom
post #24 of 136
Quote:
Originally Posted by thomasl View Post


So, it sounds like the Red slider component of Red only affects luminance and doesn't move red further/closer to the white point (i.e. affects saturation). And I would assume that this is true of the other two primaries as well. So, if red is oversaturated yet not leaning at all toward blue or green (just too far from D65), is it not possible to move it at all in toward D65? It makes sense that if red is undersaturated you can't move it out. If the red being displayed is the native red of the display then there is no way to make red "more" red. i.e. there is no way to move it out away from D65 nor a way to take away blue or green (affect the hue of red) since there is no blue or green to begin with in the native red.

I don't know if it is "right" to do it this way, but you could actually make red go to the ideal CIE spot starting with the oversaturated situation you describe. By adding some blue and some green red would be pulled away from it's oversaturated position. Whatever amounts of blue and green it takes to get red to the right xy coordinate might make the overall Y value of red too high, so the amount of red in red plus the amount of green and/or blue in red would have to be decreased, but in any case the red primary could be put very close to CIE spot-on with whatever luminance it is supposed to have.

These amounts of blue and green in red do not effect the blue or green primary colors, or for that matter the secondary colors.

Even in adjusting two secondary colors, say for example magenta and yellow, the amount of red in magenta can be changed without changing the amount of red in yellow or changing any xyY value associated with yellow (or any other color). It only effects magenta.

In fact adding red to the green primary is how we move the green primary closer to it's CIE spot.

Quote:
Originally Posted by thomasl View Post



For me personally, I've just tried to adjust the overall color luminance values to match my custom gamut values for each display via my lone Color control. Since I also only have a Tint control to affect the secondaries, I've actually just tried to get them to a position where the total dE is minimized for all 3. My primaries aren't that far off from the HD709 reference so there isn't a big difference between HD709 and my "custom" color gamut.

cheers,


--tom

Tint and the white slider in MCC

And you minimize dE as shown by what? Accupel calculator? (I ask that because Accupel's dE values are quite different from ColorHCFR's.)
post #25 of 136
I reported some bogus info. I hate reading misinformation here and elsewhere, even more when I am the source of it.

Yes, it looks like the LCDs have the same CMS that the plasmas have, which works very well. I mis-read the manual. I haven't actually seen one personally.

- Add equal amounts of the other two primaries to desaturate a primary color.
- Add unequal amounts of the other two primaries to adjust the Hue of a primary color.
- Adjust the amount of the primary color to change its brightness.
- Add the opposing primary to desaturate a secondary color (add blue to yellow)
- Add unequal amounts of the contributing primaries to adjust the hue of a secondary color.
- Equally adjust the amount of both contributing primary colors to adjust the secondary's brightness.
post #26 of 136
Quote:
Originally Posted by kjgarrison View Post

And you minimize dE as shown by what? Accupel calculator? (I ask that because Accupel's dE values are quite different from ColorHCFR's.)

Both exclusively use CIELUV (which I think is unfortunate, but that's another story), so they should be the same.

However, note that HCFR has some choices under Advanced, Preferences, Advanced that, if selected, will screw up its dE calculaton.

"Do not use luminance in delta E formula (v. 1)"

and,

"Use gamma reference to compute gray scale delta E"

should both be unchecked.

In fact, I don't know why they even offered these options. You MUST use luminance information to determine the dE of color and you should NOT use it when computing the dE of grayscale, otherwise you incorporate deviations from the selected gamma target as part of the dE value, which will likely grossly inflate it.
post #27 of 136
Quote:
Originally Posted by kjgarrison View Post

And you minimize dE as shown by what? Accupel calculator? (I ask that because Accupel's dE values are quite different from ColorHCFR's.)

I think Accupel and HCFR are both calculating dELUV.

Edit: Tom H. confirmed they are both using this.

Actually, I use my own color gamut calculator (not very user friendly - it's UNIX command line) that I wrote because I was interested in the math - it also calculates dELAB - 1976 and 1994 versions - in addition to dELUV. Up until now I had been minimizing the error using my custom gamut as reference but as I said, all our displays aren't that far off HD709 so the actual color and tint settings would probably come out the same/very close if I used HD709 as the reference. When doing measurements, I actually will look at HCFR's numbers compared to the HD709 reference since it is faster than sticking data into my program and running it. I usually only do that until after I've saved a set of measurements.

cheers,


--tom
post #28 of 136
Quote:
Originally Posted by TomHuffman View Post

- Add equal amounts of the other two primaries to desaturate a primary color.
- Add unequal amounts of the other two primaries to adjust the Hue of a primary color.
- Adjust the amount of the primary color to change its brightness.
- Add the opposing primary to desaturate a secondary color (add blue to yellow)
- Add unequal amounts of the contributing primaries to adjust the hue of a secondary color.
- Equally adjust the amount of both contributing primary colors to adjust the secondary's brightness.

Thanks much Tom for this description - it all makes sense now. Hopefully some day I'll get to try one of these displays out.


cheers,


--tom
post #29 of 136
Quote:
Originally Posted by TomHuffman View Post

I reported some bogus info. I hate reading misinformation here and elsewhere, even more when I am the source of it.

Yes, it looks like the LCDs have the same CMS that the plasmas have, which works very well. I mis-read the manual. I haven't actually seen one personally.

- Add equal amounts of the other two primaries to desaturate a primary color.
- Add unequal amounts of the other two primaries to adjust the Hue of a primary color.
- Adjust the amount of the primary color to change its brightness.
- Add the opposing primary to desaturate a secondary color (add blue to yellow)
- Add unequal amounts of the contributing primaries to adjust the hue of a secondary color.
- Equally adjust the amount of both contributing primary colors to adjust the secondary's brightness.

Concise!

As opposed to my diatribe
post #30 of 136
Quote:
Originally Posted by TomHuffman View Post

Both exclusively use CIELUV (which I think is unfortunate, but that's another story), so they should be the same.

However, note that HCFR has some choices under Advanced, Preferences, Advanced that, if selected, will screw up its dE calculaton.

"Do not use luminance in delta E formula (v. 1)"

and,

"Use gamma reference to compute gray scale delta E"

should both be unchecked.

In fact, I don't know why they even offered these options. You MUST use luminance information to determine the dE of color and you should NOT use it when computing the dE of grayscale, otherwise you incorporate deviations from the selected gamma target as part of the dE value, which will likely grossly inflate it.

Thank you Tom!

I checked and these are both unchecked.


You're (of course) correct that the dEs are the same (very close actually, not exactly). Isn't it true that the dE values are based on the ideal CIE?

Then Accupel lets you input your actual primary colors' xyY values and then it gives you dE values based on what adjustments you need to make with:
1. Luminance of the primary colors
2. Luminance and hue/saturation of the secondaries

Here are screenies. Accupel 'before' and 'after' refer to before and after doing the Copy^ for all my colors xyY data.

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.

Run5 Accupel before is with the same data as the above and ~same dEs

Run5 Accupel after is with the actual xyY values copied to the top of Accupel and the resulting dE values based on my non-spec CIE

Run5 ColorHCFRcie is the CIE. I don't have the original CIE to show, but green was off to the left (lower x).
LL
LL
LL
LL
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › Display Calibration Calculator