AVS › AVS Forum › Display Devices › Display Calibration › The Official ChromaPure thread
New Posts  All Forums:Forum Nav:

The Official ChromaPure thread - Page 86

post #2551 of 4355
5/16/12 at 1:34pm
Quote:
Originally Posted by TomHuffman View Post

Excellent question. The D3 is almost certainly the more accurate in reading luminance.

If the values are well above the minimum luminance range in the 150-200cd/m range, why would the Y value from the colorimeter that is using a physical filter to mirror the Y observer curve be more accurate than a spectro that measures the entire spectrum and can integrate the Y curve precisely?

That 3.3% difference is likely the i1 Display Pro's filter set inability to correctly match the observers Y curve. When their is plenty of light to measure I would trust a spectro over a colorimeter for Y.
Reply
Reply

AVS Top Picks

post #2552 of 4355
5/16/12 at 2:05pm
Quote:
Originally Posted by sotti View Post

If the values are well above the minimum luminance range in the 150-200cd/m range, why would the Y value from the colorimeter that is using a physical filter to mirror the Y observer curve be more accurate than a spectro that measures the entire spectrum and can integrate the Y curve precisely?

That 3.3% difference is likely the i1 Display Pro's filter set inability to correctly match the observers Y curve. When their is plenty of light to measure I would trust a spectro over a colorimeter for Y.

The 3.3% offset is present all the way up to 120 cd/m^2, so I have one vote for the D3 and one for the i1pro (note, this is the new model 2). I agree with Tom in that it really doesn't matter which is correct since relative Y is fine for all the calculations (except bt.1886, but it's a weak adjustment there). I was just curious as an academic matter.
Reply
Reply
post #2553 of 4355
5/16/12 at 2:12pm
Quote:
Originally Posted by zoyd View Post

The 3.3% offset is present all the way up to 120 cd/m^2, so I have one vote for the D3 and one for the i1pro (note, this is the new model 2). I agree with Tom in that it really doesn't matter which is correct since relative Y is fine for all the calculations (except bt.1886, but it's a weak adjustment there). I was just curious as an academic matter.

which one reads peak white higher: the D3 or the i1Pro 2?
Reply
Reply
post #2554 of 4355
5/16/12 at 2:13pm
Quote:
Originally Posted by TomHuffman View Post

This is just what you would expect. Gamma, dE color calculations, and contrast calculations all use relative luminance (grayscale doesn't use luminance at all). None of these would be affected by a Y correction. The only thing affected is absolute luminance, and this is used only for measurement of peak output and absolute black level, at least that's all I can think of that it would be used for.

Yes, that makes perfect sense.
Reply
Reply
post #2555 of 4355
5/16/12 at 2:29pm
Quote:
Originally Posted by sotti View Post

If the values are well above the minimum luminance range in the 150-200cd/m range, why would the Y value from the colorimeter that is using a physical filter to mirror the Y observer curve be more accurate than a spectro that measures the entire spectrum and can integrate the Y curve precisely?

Because a 10nm i1Pro does not precisely "measures the entire spectrum" and hence will not "integrate the Y curve precisely".
Reply
Reply
post #2556 of 4355
5/16/12 at 2:32pm
Quote:
Originally Posted by PlasmaPZ80U View Post

which one reads peak white higher: the D3 or the i1Pro 2?

Nevermind. Before going into a deep dive on interprobe comparisons I decided to do a quick check to verify. The i1pro does read higher than the D3 by 3% but only using the Argyll driver's "hires" mode (3.33 nm sampling, extended range). Switched back to the normal mode and the two probes agree to 0.1%
Reply
Reply
post #2557 of 4355
5/16/12 at 4:42pm
Here's one data point. I don't want to make too much out of it, since I only used one sample on one display (a Samsung LED flat panel).

In any case, comparing a D3 we have on hand for testing against our reference spectro, this D3's reading of peak output is off by less than 1%.

x y Y
0.304 0.320 113.1
0.304 0.320 113.0
0.304 0.320 113.1
0.304 0.320 113.1
0.304 0.320 113.0
0.297 0.324 112.1
0.297 0.324 112.6
0.297 0.324 112.1
0.297 0.324 112.6
0.297 0.324 112.5

The first 5 readings are with a Jeti spectro and the second 5 are with a D3. As you can see, the xy values vary by 0.007 in the x-axis and 0.004 in the y-axis, which is fairly common for this probe. However the average Y value for the Jeti is 113.1 and the average for the D3 is 112.4. That's a discrepancy of 0.6%.
Reply
Reply
post #2558 of 4355
5/16/12 at 5:18pm
Very interesting. I went back and did a more systematic comparison with very similar results, D3 was less the i1pro 2 by 0.7% @119 cd/m^2 and shifted -0.0078 in x, -.0012 in y (9 sample average, plasma @white 0.3132
0.3287)

dx dy dY[cd/m^2]
mean -0.0078 -0.0012 -0.95
stdev 0.0004 0.0005 0.27
Reply
Reply
post #2559 of 4355
5/16/12 at 5:51pm
The D3 is a nice meter for taking grey scale measurements off a screen. My old RS1 never looked better. Best shadow detail. Even my Chroma 5 struggled with 10 and 20 ire off the screen. Amazing meter IMHO! Fast readings too!

Hey Tom! I am using an old IBM laptop with 1.4GHz Pentium and auto Cal works great with Chromapure and the D3. Fast too!
Reply
Reply
post #2560 of 4355
5/17/12 at 10:38am
I am novice at calibration, attempting for the first time to calibrate my VT30 plasma using a Display 3 PRO and Chromapure. I have some very basic questions:
1. Should the meter touch the TV screen or should there be a distance of an inch or 2 between the meter and the TV?
2. Should the measurements be done in complete darkness or can there be ambient light?
3. I am going to use AVS HD 709 to generate the patterns. I don't have it burned on to a blu-ray, but am planning to copy it to a USB stick and plug it into my PS3 and use the PS3 as a source to display the patterns. I thought it may be better than generating the patterns from my laptop. Should that be fine?

Thanks!
Reply
Reply
post #2561 of 4355
5/17/12 at 10:46am
Quote:
Originally Posted by sachinj View Post

I am novice at calibration, attempting for the first time to calibrate my VT30 plasma using a Display 3 PRO and Chromapure. I have some very basic questions:
1. Should the meter touch the TV screen or should there be a distance of an inch or 2 between the meter and the TV?
2. Should the measurements be done in complete darkness or can there be ambient light?
3. I am going to use AVS HD 709 to generate the patterns. I don't have it burned on to a blu-ray, but am planning to copy it to a USB stick and plug it into my PS3 and use the PS3 as a source to display the patterns. I thought it may be better than generating the patterns from my laptop. Should that be fine?

Thanks!

1) Yes, you can use the meter in contact mode.

2) If you use the meter in contact mode, ambient light should not be an issue as long as the meter is flush with the display.

3) Burn the AVCHD version to a DVD and put the DVD into the PS3 (it will play like a BD) if you don't have a BD burner. Using a USB stick won't provide the same results if you plan on watching DVDs and BDs on the PS3 with this calibration (it will only be valid for watching video/movies on the USB input).
Reply
Reply
post #2562 of 4355
5/17/12 at 11:10am
Quote:
Originally Posted by PlasmaPZ80U View Post

1) Yes, you can use the meter in contact mode.

2) If you use the meter in contact mode, ambient light should not be an issue as long as the meter is flush with the display.

3) Burn the AVCHD version to a DVD and put the DVD into the PS3 (it will play like a BD) if you don't have a BD burner. Using a USB stick won't provide the same results if you plan on watching DVDs and BDs on the PS3 with this calibration (it will only be valid for watching video/movies on the USB input).

That is very helpful info. Thanks. I don't have access to a BD burner, but can burn a DVD. Would using a DVD instead of a blu-ray make a difference in the accuracy of the readings?
Reply
Reply
post #2563 of 4355
5/17/12 at 11:58am
Quote:
Originally Posted by sotti View Post

I had a customer inquire about how to import their chromapure profile so I thought I'd try to brush up on how things worked on the other side of the fence.

The quality/accuracy of technical communication (and the tone) in your posts (especially in comparison to Tom's!) reminds me why I switched from Calman to Chromapure without a glance in the rear-view mirror!
Reply
Reply
post #2564 of 4355
5/17/12 at 12:40pm
Quote:
Originally Posted by robobob View Post

The quality/accuracy of technical communication (and the tone) in your posts (especially in comparison to Tom's!) reminds me why I switched from Calman to Chromapure without a glance in the rear-view mirror!

I appreciate the feedback. I do my best to try and take care of customers. In order to be able to provide my customer with a good answer I needed a better understanding of his problem.

I didn't intend for my tone to come off as unprofessional. I was just trying to understand how the profiling works when creating profiles without a measured Y values significantly differs form the industry standard four color matrix procedure http://www.jeti.com/cms/index.php/ot...us-instruments. The ratio of Y difference for the measured value to the reference value can differ from primary to primary, and is a key component of what the matrix is trying to compensate for.

Once I understood that was his method I was apply to provide an solution to my customer. I apologize if you felt my behavior disturbed this thread.
Reply
Reply
post #2565 of 4355
5/17/12 at 2:07pm
Quote:
Originally Posted by sachinj View Post

That is very helpful info. Thanks. I don't have access to a BD burner, but can burn a DVD. Would using a DVD instead of a blu-ray make a difference in the accuracy of the readings?

no, it will behave just like a BD thanks to the AVCHD format used
Reply
Reply
post #2566 of 4355
5/17/12 at 2:48pm
Quote:
Originally Posted by sotti View Post

I didn't intend for my tone to come off as unprofessional. I was just trying to understand how the profiling works when creating profiles without a measured Y values significantly differs form the industry standard four color matrix procedure http://www.jeti.com/cms/index.php/ot...us-instruments. The ratio of Y difference for the measured value to the reference value can differ from primary to primary, and is a key component of what the matrix is trying to compensate for.

I think that it is quite possible, and in some cases even desirable, for us to have technical debates without it becoming unpleasant or unprofessional. For the record, your posts in this area have generally met this standard, something I would not say about all of your colleagues. It is easy for business interests to get in the way of a full, open, and honest exchange of ideas.

Having said that, the link you reference is to a page on the JETI web site. On that page, they write:

This is done with matrix methods and based on the fact that the color of such sources is produced by a superposition of the three primaries (RGB). An easy procedure is described in [1]. A better approach is shown in [2].

What is [2], the "better approach" they refer to? It is a paper by NIST's Ohno and Hardis entitled "Four-Color Matrix Method for Correction of Tristimulus Colorimeters".

This paper is exactly the same one I referenced below and the just the method we use. I posted a copy of it on our web site. That method does not use luminance (Y) data. Quoting from it: "A new technique for the matrix method (named Four-Color Method) has been developed, which is based on the (x, y) values only, and is independent of Y value."

Having said that, whether one bases a matrix correction on xyz without luminance or on XYZ which uses luminance is, I think, of very little consequence for the reason I have stated previously. Properly functioning colorimeters already do an excellent job of reading luminance, so any correction made here would be well below the threshold of visibility.

Furthermore, even if a larger correction were needed, since relative luminance is by far the more important consideration, small errors in absolute luminance wouldn't matter much to the calibration results in any case.
Reply
Reply
post #2567 of 4355
5/17/12 at 2:58pm
Quote:
Originally Posted by PlasmaPZ80U View Post

no, it will behave just like a BD thanks to the AVCHD format used

Thanks for your help. Can't wait to start the calibration exercise
Reply
Reply
post #2568 of 4355
5/17/12 at 4:17pm
Quote:
Originally Posted by TomHuffman View Post

This paper is exactly the same one I referenced below and the just the method we use. I posted a copy of it on our web site. That method does not use luminance (Y) data. Quoting from it: "A new technique for the matrix method (named Four-Color Method) has been developed, which is based on the (x, y) values only, and is independent of Y value."

Four-Color Matrix Method for Correction of Tristimulus Colorimeters - Part 2

The performance of the Four-Color Method has been
studied further for spectral variations of displays
including CRTs, LCDs, and OLED displays. A theory for
additional photometric (Y) correction is added.
Reply
Reply
post #2569 of 4355
5/17/12 at 7:02pm
Quote:
Originally Posted by ConnecTEDDD View Post

Four-Color Matrix Method for Correction of Tristimulus Colorimeters – Part 2

The performance of the Four-Color Method has been
studied further for spectral variations of displays
including CRTs, LCDs, and OLED displays. A theory for
additional photometric (Y) correction is added.

Yes, there was a subsequent paper published a year later in which they showed how Y correction could be added. However, that was not the paper referenced on the Jeti page and described here as the "industry standard four color matrix procedure."
Reply
Reply
post #2570 of 4355
5/17/12 at 7:05pm
OK, the mysterious problem with the 75% of Rec. 709 gamut has been fixed. It can be downloaded from the 2.2.13 link on the ChromaPure News page.

This was really difficult to track down. The problem only arose if you had never selected a signal generator. If you had previously used a signal generator--even if you were not using one currently--the problem would not appear. This is why my tests failed to reveal the issue.
Reply
Reply
post #2571 of 4355
5/17/12 at 7:13pm
Tom, do you ever sleep????
Reply
Reply
post #2572 of 4355
5/17/12 at 7:56pm
Quote:
Originally Posted by buzzard767 View Post

Tom, do you ever sleep????

I am a night person. I also sleep late.
Reply
Reply
post #2573 of 4355
5/18/12 at 6:38am
Just got my Duo back from repair, so had chance to see how much the latest Chromapure autoCal faired, now normally I would fine tune after being run, but I felt no need this time.

Has there been much improvement with AutoCal, or did I just hit lucky?
Reply
Reply
post #2574 of 4355
5/18/12 at 9:30am
Quote:
Originally Posted by magicj1 View Post

Just got my Duo back from repair, so had chance to see how much the latest Chromapure autoCal faired, now normally I would fine tune after being run, but I felt no need this time.

Has there been much improvement with AutoCal, or did I just hit lucky?

I don't know but I was very impressed with how well it worked on my old 1.4GHz P4 laptop with 1GB memory. Flawless and fast using the D3 pro meter with my DVDO Duo.
Reply
Reply
post #2575 of 4355
5/18/12 at 5:39pm
Quote:
Originally Posted by TomHuffman View Post

OK, the mysterious problem with the 75% of Rec. 709 gamut has been fixed. It can be downloaded from the 2.2.13 link on the ChromaPure News page.

This was really difficult to track down. The problem only arose if you had never selected a signal generator. If you had previously used a signal generator--even if you were not using one currently--the problem would not appear. This is why my tests failed to reveal the issue.

It certainly fixed it for me. Thanks Tom.
Reply
Reply
post #2576 of 4355
6/1/12 at 7:58am
I was attempting to use the following guide

http://www.avsforum.com/avs-vb/showthread.php?t=1134710

to calibrate my epson 8350 using Chromapure.

In the section about adjusting Color Gamut, he has the following tip found below. Are these % bars available in Chromapure? I can't seem to find them.

I've been using the Advance CMS and just calibrate x & y for each color at 75% until in lands on the 75% mark on the Gamut chart. However on a few colors like green, adjusting hue or sat moves both x & y.

Special Hint!
I have found a very easy way to hit the correct color locations. This takes away the headache of trying to hit x, y locations (as shown above) and how you adjust saturation and hue to get there. See the red, green, and blue bars in the HCFR picture above? (three pictures back) In the picture, the bars show 98%, 100%, 100% to the left of the ftL and cd/m2 readings. These are the bars I'm talking about. Here are the target %'s for the red, green, and blue bar for each 75% color saturation location. (They are also listed in my Excel spreadsheet on the "Calibration Aid" tab as shown a couple of pictures back).


Code:
Red Bar Green Bar Blue Bar
Red Primary 378% 24% 24%
Green Primary 15% 133% 15%
Blue Primary 64% 64% 555%
Yellow Secondary 106% 106% 17%
Cyan Secondary 24% 120% 120%
Magenta Secondary 247% 41% 247%For example, lets say you want to calibrate red. To hit the correct saturation, all you have to do is increase or decrease the red saturation slider in the Epson RGBCMY menu until the red bar reads 378%, and the Green and Blue bar read 24%. If the green and blue bars are not equal, use hue to correct this. Hue is always used to balance colors. When calibrating blue for example, you want green and red to be equal. This creates the correct hue. When adjusting a secondary color, like yellow for example, you want an equal balance of green and red. (red and green will have a high %, while blue will be low.)

Keep in mind that you still need to set brightness (Y) for red, and all the colors The target value will be shown in the "Calibration Aid" spreadheet, or you can calculate it based on the 100% gray window. Remember, calibrate red, green, and blue before you calibrate the secondary colors.
Reply
Reply
post #2577 of 4355
6/1/12 at 8:06am
There is a much simpler way to this in ChromaPure. In the standard Color Management module, just select "75% of Rec. 709" as your target gamut and then calibrate normally.

Be sure to download and install the latest build available on the CP News site.


Quote:
Originally Posted by jmoakk View Post

I was attempting to use the following guide

http://www.avsforum.com/avs-vb/showthread.php?t=1134710

to calibrate my epson 8350 using Chromapure.

In the section about adjusting Color Gamut, he has the following tip found below. Are these % bars available in Chromapure? I can't seem to find them.

I've been using the Advance CMS and just calibrate x & y for each color at 75% until in lands on the 75% mark on the Gamut chart. However on a few colors like green, adjusting hue or sat moves both x & y.

Special Hint!
I have found a very easy way to hit the correct color locations. This takes away the headache of trying to hit x, y locations (as shown above) and how you adjust saturation and hue to get there. See the red, green, and blue bars in the HCFR picture above? (three pictures back) In the picture, the bars show 98%, 100%, 100% to the left of the ftL and cd/m2 readings. These are the bars I'm talking about. Here are the target %'s for the red, green, and blue bar for each 75% color saturation location. (They are also listed in my Excel spreadsheet on the "Calibration Aid" tab as shown a couple of pictures back).


Code:
Red Bar Green Bar Blue Bar
Red Primary 378% 24% 24%
Green Primary 15% 133% 15%
Blue Primary 64% 64% 555%
Yellow Secondary 106% 106% 17%
Cyan Secondary 24% 120% 120%
Magenta Secondary 247% 41% 247%For example, lets say you want to calibrate red. To hit the correct saturation, all you have to do is increase or decrease the red saturation slider in the Epson RGBCMY menu until the red bar reads 378%, and the Green and Blue bar read 24%. If the green and blue bars are not equal, use hue to correct this. Hue is always used to balance colors. When calibrating blue for example, you want green and red to be equal. This creates the correct hue. When adjusting a secondary color, like yellow for example, you want an equal balance of green and red. (red and green will have a high %, while blue will be low.)

Keep in mind that you still need to set brightness (Y) for red, and all the colors The target value will be shown in the "Calibration Aid" spreadheet, or you can calculate it based on the 100% gray window. Remember, calibrate red, green, and blue before you calibrate the secondary colors.
Reply
Reply
post #2578 of 4355
6/1/12 at 8:27am
Quote:
Originally Posted by TomHuffman View Post

There is a much simpler way to this in ChromaPure. In the standard Color Management module, just select "75% of Rec. 709" as your target gamut and then calibrate normally.

Be sure to download and install the latest build available on the CP News site.

Thanks, I just downloaded the latest version! So just follow the ChromaPure guide for standard cms?
Reply
Reply
post #2579 of 4355
6/10/12 at 3:14pm
I have had a chance to spend some time with the new i1Pro 2, and I wanted to report my findings.

The accuracy of the unit is the same for the original i1Pro. It looks and feels like it is better built, and it is certainly heavier. It is better isolated from temperature and it reads higher levels of luminance, but the vast majority of the improvements were for reflectance readings for those who use it for photography and pre-press work, not for the emissive mode that is used for display measurements.

As before, the included diffuser is not color neutral, so for accurate readings from the lens of a projector, you should first correct the meter's response using the i1Pro in Standard mode reading off the screen. Once corrected, lens readings should be quite accurate.

Speaking of accuracy, the i1Pro does best reading off of screens that receive light from a UHP bulb. It does less well with LED illumination and on flat panels generally.

It is critical that upon initial calibration that you lay it face down on a flat surface, preferably the included calibration plate. After that you can take periodic dark readings by covering the sensor in whatever way that is most convenient.
Reply
Reply
post #2580 of 4355
6/10/12 at 3:17pm
Quote:
Originally Posted by jmoakk View Post


Thanks, I just downloaded the latest version! So just follow the ChromaPure guide for standard cms?
Yes. Once you select that gamut, all you have to concern yourself with is using the right test patterns (CP provides built-in patterns that are easiest to use if your PC or laptop has an HDMI out.
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration

AVS Top Picks

AVS › AVS Forum › Display Devices › Display Calibration › The Official ChromaPure thread