The Official ChromaPure thread - Page 81

This is really odd. I just cannot reproduce this.

i1Pro Grayscale



D3 Native Grayscale



D3-i1Pro Corrected Grayscale



This particular D3 reads considerably too low in red, so the correction is easy to see. The post-correction results are within 0.001-0.002 of the i1Pro.

I am stumped. The only thing I can suggest is to reinstall .Net.

Is anyone else seeing this?

this may have something to do with the display clipping white ??

One more step Tom, after finishing what you did can you re-measure a grayscale using the i1Pro AGAIN?

I will do another test on another computer and monitor just to see if this repeats. Will post in a couple hours.

I suspect if possible it may be a USB thing.......

You couldn't do this, because then you would be applying the correction the i1Pro.
1. Initialize the i1Pro, and take reference measurements in the Meter Correction module.
2. Click Disconnect in the Meter module and physically unplug the i1Pro. You are now done with this meter for this session.
3. Initialize the D3 or any other colorimeter, and take field measurements in the Meter Correction module.
4. Click Apply.
5. Use the colorimeter for the remainder of all measurements.

Ok, another computer, another installation of the latest CP again and another monitor (Acer H243H) instead of the Viewsonic that I used.

First, D3 baseline


Then, an i1Pro baseline


Now, the D3 after profiling against the i1Pro. Is this acceptable results after profiling?


The i1Pro by itself, the reading has always been consistent but that does not apply its reading is correct. Nonetheless, consistent to its own. The readings above 30IRE is very stable.

The D3 has been very consistent through out and never experience any problem. The thing is that even after the profiling against the i1Pro, as you can see the numbers are adjusted but not to the i1Pro baseline. I have done this to two plasma, two projectors and three different computers and it has always been like this.

Now another funny thing is look at how the i1Pro measures AFTER being profilied by the D3. i1Pro is the reference and D3 is the field meter. The CP profiling actually changed how the reference measures! Red increases and Blue decreases.


And here is the calibration table for your reference.


Lets try to be logical, even if my i1Pro is a lemon, at least its measurements are consistent. So why the profiled-D3 does not produce the same result? I mean I have tried this combo on so many displays and two different screens. And why would the reference meters measurements changes dramatically after being profiled to? This part may be a bug in CP I am not sure but that is the logical thinking since both meters are consistent on their own.

Another funny thing about CP is that even if you don't disconnect the i1Pro, you can still measure with the D3 but of course the measurements are erratic.

The questions are;

1. Am I seeing a bug of the CP?

2. Do I have a bad i1Pro?

3. Am I not following procedures? Here is what I did;
- Plug in both meters.
- Run CP.
- Initiate i1Pro.
- Take a grayscale measure.
- Take the reference reading in Meter Correction.
- Disconnect the i1Pro.
- Initiate the D3.
- Take a grayscale measure with the D3.
- Take the field reading in Meter Correction.
- Click 'Apply'.
- Take a grayscale measure with the D3 again.
- Disconnect the D3.
- Initiate the i1Pro.
- Take a grayscale measure with the i1Pro (after the D3 has been profiled).

4. On different displays, the profiled D3 result can vary and be more far off from the i1Pro baseline than the others. Is that true?

Anyone having such problems?

Testing on different displays is not helpful. The application does not know nor care what display you are using. One problem with this latest test is that the i1Pro and the D3 are so close to one another to begin with it is hard to see what is happening.

My sense is that you are making this much more complicated than it needs to be.
1. There is no reason to take a D3 baseline grayscale.
2. Avoid having both meters plugged in at the same time. This may cause some driver interaction.
3. You CANNOT go back and use the i1Pro again after having used it as a reference to profile another meter. When you do this you are applying a correction to the reference meter based on a comparison between the reference and field meter.

Please try these steps exactly. Start with the D3 unplugged.
1. Initialize the i1Pro, take a dark reading, and then take a baseline grayscale reading.
2. Take reference measurements in the Meter Correction module.
3. Click Disconnect in the Meter module and physically unplug the i1Pro. You are now done with this meter for this session. Don't use it again unless you start a new session with no corrections.
4. Plug in and initialize the D3, and take field measurements in the Meter Correction module.
4. Click Apply.
5. Take grayscale measurements with the D3.

Compare the two grayscale runs.


For the reason I have already explained. You are comparing:
1. The native i1Pro response
with
2. The i1Pro response corrected by a profile intended for the D3.

Once the corrections are created, they will apply to any measurement taken during that calibration session using any meter.

I followed your instructions and actually repeated it twice to make sure that it is what it is.

i1Pro Baseline.


D3 after profiling to the i1Pro.


And the Correction table for your reference.


I really hope that was it but no. It just seems like the correction numbers are not fully added to the D3....... I know it sounds crazy but.

A note, I have made adjustments to the monitor itself and that's why the numbers doesn't look like what it was before this post.

Tom,

I have followed your instructions 100%. Does CP check the versions of all supporting softwares? I think I have the latest .Net on all my computers. Now that I think about it I actually installed the CP on 3 different desktop and 1 laptop and they all seems to run fine but this has happened to all.

Yes, this time I physically unplugged the i1Pro and there is only one meter connected at a time.

From the results above, unplugging help but doesn't solve the mystery. The blues are now very very close to say the least but the red is still lacking.

I suppose CP just add the differences from the 100% reading and just that particular amount linerly through all reading of the field meter right? If so, due to the non-linearity of either meter the field meter can still be wrong starting from low IRE. Is that true?

Would a two point profiling help if the above is true? Just like how we adjust grayscales with two points. The reference will take a reading at 40IRE and 100IRE and the field meter would do the same then CP will compensate the difference?

I think that your expectations are too high. If you look closely at the numbers, it looks quite good until you get to 30% and below. I would attribute this to the low-light deficiencies of the i1Pro. Otherwise, this result is exactly what I would expect. This is precise why you use profiling--to get the benefits of a colorimeter at low light levels with the accuracy of a spectro.

% Stim	i1Pro		D3		Difference
	x	y	x	y	x	y
10	0.289	0.320	0.288	0.313	-0.001	-0.007
20	0.310	0.336	0.306	0.336	-0.004	0.000
30	0.319	0.343	0.314	0.343	-0.005	0.000
40	0.320	0.342	0.317	0.341	-0.003	-0.001
50	0.318	0.339	0.316	0.340	-0.002	0.001
60	0.315	0.335	0.313	0.335	-0.002	0.000
70	0.312	0.331	0.311	0.332	-0.001	0.001
80	0.311	0.330	0.310	0.330	-0.001	0.000
90	0.310	0.327	0.309	0.327	-0.001	0.000
100	0.312	0.331	0.311	0.330	-0.001	-0.001
					0.002	0.001	Ave.

I just don't see anything in these results that would concern me.

Thank you, Tom. If this is common then at least I know my meters are fine and don't need replacement.

And now I know that not to plug in more than one Xrite meters at the same time. I thought having two meters plug in is convenient because I can double check the D3 results with the i1Pro to make sure the D3 profile is correct!

One question though, the x error is 6% and y error is 3% plus the inherited error from the i1Pro say 2% from true reference. We are talk 8% in x and 5% in y. So is it fair to say that even after calibration, our display will always be 5-10% off the true reference? Imagine 5% more red will be quite visible even to the human eyes.

Thanks again.

No, I wouldn't say that at all. I don't know where you get these figures. Once the colorimeter is profiled to the reference at adequate light levels, then it should be accurate to within 0.002 of that reference at all light levels. In this case the reference--the i1Pro--may not maintain its accuracy as well at low light levels.

Also, even assuming those error rates are valid, the conclusion is not. Because errors are random, you can't assume that all errors are perfectly additive. Two 2% errors are just as likely to cancel each other out as they are to result in a 4% overall error. You also should not assume that a 2% error in x and y is twice as inaccurate as a 4% error in either alone. For example, 0.3167, 0.329 is more inaccurate than 0.3147, 0.331.

Thanks for the explanation. Very helpful.

I generally go by the deltaE rather than obsess about hitting 0.313/0.329 exactly. If it's below 1.5 or 2%, then I'm good.

This is a perfectly vaid way to go about it, isn't it?

Yes, dE is the easiest and most straightforward way to determine error. The only complication with this is knowing how to interpret the values while keeping in mind that there are different dE formulas that yield different results.

I use CIE94 and go for under 1.5% in Grayscale and Colors.

I found an easy way to add a Measure All feature to the Advanced color Management module, so I folded this into a new build, along with the previously-announced enhancements to the AccuPel HDG-4000 rs-232 support, and some minor enhancements to the Detailed Calibration Report.

A download link and Release Notes are available on the News page on the ChromaPure web site.

Note: Don't click the "Measure All" button in the Advanced Color Management module until the meter is initialized and the test pattern generator is setup and working. Also, don't touch the mouse or any keys until the measurements have finished. It takes about 90 seconds for all of the saturation and amplitude readings to complete.

To use the Advanced Color Management Module, you need to setup a supported signal generator. These include:

• AccuPel DVG-5000
• AccuPel HDG-4000
• Lumagen Radiance
• Internal Test Patterns

Thanks Tom, this is much appreciated, looking forward to testing this

OK, you figured it out. Don't click "Reset Data" first. That unselects the white radio button.

Yes, I guess I did this to make sure the pattern/meter were initialised.

If it's easy to implement, it would be great to have a "measure saturations" and a "measure amplitudes" button under the "measure all", so we can split down depending on what we're working on.

Thanks again Tom!

With CP 2.1.12 & 2.1.13, I couldn't reset the data of 'Post Calibration Color Gamut' so I had to close the tab to measure again. But it didn't show the luminance chart before I started to measure the 40 IRE or higher. Simply it was blank..

Nobody experienced this?

I am not sure what you mean. There is no IRE selector in the Color Gamut modules.

I could not reproduce the reset problem. I can only suggest that you reinstall or repair .Net.

My mistake! Luminance chart for RGBCMY was empty when re-opened the 'Post-calibration color gamut' and it's not shown again until measured B, Y, C &M.
I can upload the video in youtube if needed.

Yes please.

Added my video in Youtube.

http://youtu.be/A2EboPvE5ro

1. Reset doesn't work
2. Luminance chart does not show
3. closing the color gamut tab does not reset the chart

I cannot reproduce the first two problems you are seeing, and we have done nothing in recent builds to change this in any case.

http://www.chromapure.com/bugtest/reset/reset.html

Either there is a problem with your installation (try reinstalling .Net and then downloading and reinstalling CP again) or there is some other application running on your PC that is interfering with CP, such anti-virus, etc.

The third issue is not a problem. This is by design. The data in the pre/post-calibration grayscale and color gamut modules are persistent. They will not go away unless you close the application and start a new calibration session or measure over them.

Hi Tom, I have recently replaced my beloved JVC RS40 for a new SONY VPL VW90ES SXRD projector. I thought its time for a change after two straight years of DiLA user. Was so happy with the new acquisition and immediately I fired up the Chromapure calibration software to perform the essential tweaks using the auto-calibration feature built-in. I am a happy user of the auto-cal feature back then when I am still using the JVC. I have performed the following steps before the auto-cal which I have did for my JVC projector in the past and all went smooth until NOW...

Hardware used in my video chain
- DVDO iScan Duo - using the built-in signal pattern generator - with Chromapure (latest version - 2.2.13)
- Chroma 5 colorimeter (takes quite long to read )

Here are the steps I did...
1) Perform a calibration on the contrast, brightness using the pluge pattern
2) Perform a white balance calibration using Chromapure (under calibration section) for better accuracy (less chore for auto-cal)

Projector settings as follows (not sure if you require it but better to provide)

Picture Mode: User

Cinema Black Pro:
- Advanced IRIS: Off (this is to prevent any unnecessary auto adjustment in IRIS which may affect the results)
- Lamp Control: Low

- Motion Flow:
- Both Front Projection and Motion Enhance set to OFF (rationale same as IRIS)

Contrast: 90 (default no change)
Brightness: 46 (default value is 50 - adjusted using Pluge pattern)
Color: 48 (default is 50 - adjusted using the color decoding feature under the Calibration secton)
Hue: 50 (default no change)
Sharpness: 10 (default no change)

Color Temp: Custom 4
- Gains RED = 16
- Gain GREEN = 0
- Gain Blue = -10

- Bias RED = 5
- Bias GREEN = 0
- Bias BLUE = -3

Expert Settings:
Gamma = 9
x.v. Color: OFF
Colorspace: Normal (to coincide with Rec HD709)
RCP: OFF

The pre-grayscale tracking result showed alot of blue and too little red. So the adjustment under 80% grayscale pattern showed the abovementioned results.

Below are the screenshots of the calibrated results for your viewing.











Here's where things start to get weird?!



As you can see that the gamma readings go awry AFTER 50% grayscale...I have tried many times to get it right but to no avail. Everything else is in place except the gamma values. May I know what is happening? Is there anything wrong with the latest version of the Chromapure or the problem is something else? Take note that there is NO CHANGE in any of the hardware components in my video chain EXCEPT I upgraded to the latest version of Chromapure i.e. 2.2.13). This has never happen to my JVC projector before? What gives?

Let me know if you need more data...

BTW the picture looks great even though the gamma looks 'bad'...Is this normal I wonder?!

The gamma is terrible in the Before reading. The calibration gets it right until 50 where it runs out of steam. It looks to me as if you are driving the unit too hard and clipping the whites. See that 100% is no brighter that 90% Trying turning contrast down and/or using different gamma presets so you can start from a better baseline.

Thanks Tom. You are right about the Contrast set too high. This time round, I managed to bring down the Contrast to 70 and it now provides me with a smooth gamma curve at an average of 2.20.



I also re-do my other settings...but this time round, I find all the colors are under-saturated. But the grayscale is near perfect! Do you think I should be overly concern about the under-saturation? Do you think I should use Wide gamut? Or perhaps increase the color to dial the primaries (RGB) near to the reference points?

Changed the colorspace to Wide 1 to increase the color gamut and resolved the under-saturation issue with all the colors.