i1 display 3 profiled to i1 pro 2: some numbers - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 25 Old 03-26-2013, 12:54 PM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
Hi everyone,

So I've profiled the i1 display 3 colorimeter to the i1 pro 2 spectrophotometer. The display 3 was 14 cm from the display. The i1 pro 2 was 24 cm from the display both when measuring and profiling (with is also measuring of course). Basically, I guessed how much I was to put the i1 pro 2 farther back:-). I went with 10 cm as you can deduct from above numbers.

Display in all cases was allowed to warm up for about 45 minutes to 1 hour.

- The first tab of the Excel sheet shows a failed attempt to create a profile. I neglected to put the i1 pro 2 farther back there. D'oh!
- The second tab shows a much better result except for a slight error for red (I'm thinking repeatibility error here). Here, I put the i1 pro 2 farther away (10 cm).
- The third tab is a verification of the profile 1 day after it had been created. It also makes a measurement run with the profile disabled.

So there you have it. Some nice data on profiling to show its use.

Test with meter profiling.xlsx 158k .xlsx file
Attached Files
File Type: xlsx Test with meter profiling.xlsx (158.1 KB, 53 views)

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
Sponsored Links
Advertisement
 
post #2 of 25 Old 03-26-2013, 01:09 PM
Senior Member
 
rmongiovi's Avatar
 
Join Date: May 2005
Posts: 478
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 26
Why the added distance?
rmongiovi is offline  
post #3 of 25 Old 03-26-2013, 01:14 PM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
The display 3 reads more of the screen than the i1 pro 2 if they are placed at the exact same distance.
So to make them read the same area, the i1 pro 2 must be placed further back.

I'm not really sure 10 cm is enough. But judging from the profile it was fine.

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #4 of 25 Old 03-26-2013, 01:57 PM
AVS Special Member
 
PlasmaPZ80U's Avatar
 
Join Date: Feb 2009
Posts: 7,052
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 187
profiling doesn't need to be too technical... you can even just place each meter in the center of the screen in contact mode either together in single-pass mode or one at a time in multi-pass mode if you are concerned about screen uniformity

in the end, as long as the profiled colorimeter reads the same (or realistically almost the same) as the spectro the profile should be valid...

also if you don't need to measure black/very dark grays/very dark patterns in general you might just opt to use the spectro alone
PlasmaPZ80U is offline  
post #5 of 25 Old 03-26-2013, 01:57 PM
Senior Member
 
<^..^>Smokey Joe's Avatar
 
Join Date: Aug 2006
Location: Christchurch, New Zealand
Posts: 368
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 39
This information may help this topic.
The core data was provided by Spectracal, I added some further data and looked at other aspects.

This is area verse distance per probe design



This is relative 50mm area verse distance



This is detail observations for i1pro verse C6/D3
Note: Where things get fuzzy is the cross over of FWHM and TLT areas.

Jeroen1000 likes this.

Masterpiece Calibration Ltd
Christchurch NZ
<^..^>Smokey Joe is offline  
post #6 of 25 Old 03-26-2013, 02:03 PM
AVS Special Member
 
PlasmaPZ80U's Avatar
 
Join Date: Feb 2009
Posts: 7,052
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 187
http://www.tlvexp.ca/2012/04/do-calibration-tables-really-work-for-tri-stim-devices/

if you read the addendum on varying distance for the jeti in this article, you'll see the whole FOV argument is greatly exaggerated on a typical display with decent screen uniformity
PlasmaPZ80U is offline  
post #7 of 25 Old 03-26-2013, 04:03 PM
Senior Member
 
<^..^>Smokey Joe's Avatar
 
Join Date: Aug 2006
Location: Christchurch, New Zealand
Posts: 368
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 39
It must be interpretation of results and of the testing methods.

The addendum, in my interpretation only proves that the Jeti reports different results to the C6 in contact mode.
The jeti has a 2deg feild of view, the movements and measures would largely be similar each time, the difference between the 40inch distance(1016mm) and 85inch(2159mm) is equal to an area difference of roughly 32mm to 68mm
One of the reasons guidelines for measurement stimpulates narrow field of view (<2deg)is to lower contamination of source signal from off angle shifts and reflective light contamination.

Probes like the C6/D3 NEED to match the FOV of the reference device to account for the above and any non uniformity.
In contact mode other issues such as inbetween pixel darkness (plasmas) can interfere with exact measures.

If you don't match FOV within reason(TLT and FWHM issues aside), I wouldn't bother with an offset matrix. The C6/D3 as standard could possibly be more accurate.

Masterpiece Calibration Ltd
Christchurch NZ
<^..^>Smokey Joe is offline  
post #8 of 25 Old 03-26-2013, 04:53 PM
AVS Special Member
 
Michael TLV's Avatar
 
Join Date: Feb 2000
Location: THX/ISF Calibrationist/Instructor, Calgary, AB, Canada
Posts: 6,726
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 114
Greetings

The C6 results were based on the calibration tables in that test. The Jeti is simply showing how close those calibration tables really got when compared to a reference device. Another occasion where the tables failed or not ... depends on how you interpret it.

Regards

Michael Chen @ The Laser Video Experience
ISF/THX/TLV Video Instructor
The Video Calibration Education Hub - www.TLVEXP.com

Michael TLV is offline  
post #9 of 25 Old 03-27-2013, 12:24 AM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
Thanks for the insightful charts Joe. Some "Googling" is required to grasp the finer details of that last one biggrin.gif

@ PlasmaPZ80U. I have not double/triple checked so I can't be 100% certain, but my experience so far is that you do have to make some minimal effort to match the FOV. It may not be as critical as proclaimed here and there, but putting both meters at the same distance (+-14 cm) did not yield a good* profile on my PF11EK Pro plasma. Could be screen dependent, or not..So contact mode may not be the best idea on my screen.

*That having said, good and bad here are very relative. Just look at the dE number's for blue. The failed profile (first tab) shows a dE deviation of 0.1484 from the spectro, the good profile on the next tab shows a deviation of 0.0978. A difference of 0,0506 dE.
Funnily, on another testing run, I had already moved the i1 display 3 around at various distances (as far as 30 cm back to only 5 cm away from the screen) and moved it from left to right in the pattern box, and x,y values usually did not move by more than 0,001. I don't know how the spectro reacts to being moved.

Edit: I just noticed I made a mistake for Y on the last tab. What is an acceptable deviation for this parameter? Someone told me that you should match Y between the 2 meters before creating a profile. I did not do that but it is hard to judge whether this could become an issue without knowing what the max. deviation for Y should be.

The good profile made on the 25th March had below deviation for Y
White.....Red....Green....Blue.....Cyan....Magenta...Yellow
0,0121 0,0691 0,1200 0,1008 0,0213 0,0874 0,0688

But the verification the day after, shows a significantly(?) larger deviation:

White.....Red....Green....Blue.....Cyan....Magenta...Yellow
0,9240 0,2642 0,6482 0,1371 0,7487 0,3324 0,8480

Granted, I did not place the meters in the exact same location (but pretty close) as the day before.

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #10 of 25 Old 03-27-2013, 09:12 AM
AVS Special Member
 
PlasmaPZ80U's Avatar
 
Join Date: Feb 2009
Posts: 7,052
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 187
Quote:
Originally Posted by Michael TLV View Post

Greetings

The C6 results were based on the calibration tables in that test. The Jeti is simply showing how close those calibration tables really got when compared to a reference device. Another occasion where the tables failed or not ... depends on how you interpret it.

Regards

didn't that addendum also show that the Jeti measured the same on that particular display regardless of how close or far the meter was from the display?

and if you were to do the same check with the C6 instead, do you think varying distance would matter (given it's much larger field of view)?
PlasmaPZ80U is offline  
post #11 of 25 Old 03-27-2013, 10:29 AM
AVS Special Member
 
Michael TLV's Avatar
 
Join Date: Feb 2000
Location: THX/ISF Calibrationist/Instructor, Calgary, AB, Canada
Posts: 6,726
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 114
Greetings

In the ideal, the c6 (or similar device) should read relatively the same regardless of the distance/FOV as long as it is still looking at the same pattern ... versus the pattern and the bezel around the TV and the window behind the TV. smile.gif So if you started with a window pattern and then switched to a full frame pattern, it should still read the same (except on a plasma).

If it doesn't read the same (relatively speaking) in such a condition, doesn't that make the device rather useless? It means the only thing a person is ever calibrating is a 1 to 2 inch circle in the center of the TV and nothing else (Even though mosts sets are more uniform today than they are not). And something 5 inches to the left is wrong ... frown.gif

Regards

Michael Chen @ The Laser Video Experience
ISF/THX/TLV Video Instructor
The Video Calibration Education Hub - www.TLVEXP.com

Michael TLV is offline  
post #12 of 25 Old 03-27-2013, 02:10 PM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
Quote:
Originally Posted by Michael TLV View Post

Greetings

In the ideal, the c6 (or similar device) should read relatively the same [...]

So for x,y relatively the same is a maximum error of 0,002 - 0,003 considering the tools I'm using. But what is relatively the same for Y? I really have no idea on that.

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #13 of 25 Old 03-27-2013, 02:36 PM
AVS Special Member
 
Michael TLV's Avatar
 
Join Date: Feb 2000
Location: THX/ISF Calibrationist/Instructor, Calgary, AB, Canada
Posts: 6,726
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 114
Greetings

Actually, the real value of Y doesn't matter in a profile. Calibrators don't shoot for specific Y values. Don't really use it for setting contrast or light output. All you can do is trust that it is consistent relative to itself.

For CMS ... when you take the reading of white, you establish the relative values of all the other colors anyway.

The 3 rules for setting contrast don't exactly yell out a specific Y value.

Regards

Michael Chen @ The Laser Video Experience
ISF/THX/TLV Video Instructor
The Video Calibration Education Hub - www.TLVEXP.com

Michael TLV is offline  
post #14 of 25 Old 03-27-2013, 03:37 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,585
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 164
If little x and little y remain constant, then the absolute accuracy of big Y isn't much of an issue.

But when we talk about how the meter captures and reports data, we are talking about capturing big X, big Y and big Z. the little x,y is a relationship between all three variables.


x = X/ (X+Y+Z)
y = Y/(X+Y+Z)

So the accuracy of Y is critically important to creating good x,y numbers, but if your x,y numbers compare favorably to a different meter and only the big Y values differ, then it has more to do with sensitivity than accuracy.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #15 of 25 Old 03-28-2013, 12:05 AM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
So, If I'm getting this right, as long as big Y varies correctly (relatively speaking) when the stimulus of a pattern changes, everything is fine = accurate.

However, I'm sure I've read about this before and a large disagreement ensued, why is a profile then also correcting for Y and not x,y alone?

ps: I'm using big Y (in absolute numbers) to measure peak white from time to time:-)

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #16 of 25 Old 03-28-2013, 02:24 AM
Senior Member
 
<^..^>Smokey Joe's Avatar
 
Join Date: Aug 2006
Location: Christchurch, New Zealand
Posts: 368
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 39
As others eluded too, capital Y is a relative measure. Which means what ever the maxY is, all other calculations are based upon it.Which also means that absolute Y accuracy for the device measure is less important.

What others elude to in some respects trying to point out FOV could also be a overrated concern is for a similar reason.
The reasoning is similar to the concept of Y being relative where the offset matrix in theory takes into account the differences of FOV and xyY differences.
In simple terms, based upon a simple offset and multiplyer, or linear equation.

Thats the theory.
The problem is this, the offset matrix is a point in space, based upon one luminance.(what ever that may be).
This works as long as the relationships between the variables stay linear.

The concept of the offset matrix for colorimeters is not new and studys show the methods works, but the real world doesn't always play ball.
The variables are the problem where the measurement device, display, source and the environment can all have the potential to not return a linear result.

In the case of FOV, where you place meters at a distance.
Here you can have quite different capture angles between the reference device and final measurement device. Because of this difference there will be off axis colour drifts, especially in the TLT capture area. The environment can have different effects on each type of probe. This is realtive to the probe, but different to each other.
This is why reference devices have <2deg capture angle.

The reason some have found it can be hit and miss at times for some unknown reason is a variable has snuck in, like the difference of FOV, glare from the environment, off axis colourshift from the display etc etc.

Another example is in contact offset matrix
The viewing area is different between a colorimeter and an i1pro, with plasmas and some panels there is a viewable and measurable dark area between pixels. This can make plasmas measure a little green.Here you can have a dominant different FOV and dark area measurement and could create a non linear tracking.

The point of this dribble is to check what you do, that is, compare the results from the two probes, before a matrix is created and after and before you dive head first into a calibration.
Jeroen1000 and PlasmaPZ80U like this.

Masterpiece Calibration Ltd
Christchurch NZ
<^..^>Smokey Joe is offline  
post #17 of 25 Old 03-28-2013, 05:52 AM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
Thank you all for the excellent information.

Joe, why is display linearity an issue with profiling? Unless, I have not understood you correctly and you are *not* linking a non-linear display device to meter profiling.
I understand it can be an issue when trying to get the colors accurately calibrated: the display may not perform linear and only correcting the primaries and secondaries at a certain % saturation and % stimulus is no guarantee that they will also be correct at other saturation and stimulus points.

From my loosely (meaning, just using common sense) conducted tests, I found out that I have make sure no ambient light can enter the meter (I also dim the laptop and keep it away from the meter as good as possible), take care of the FOV to a reasonable extent and I'm good to go.

I've been tweaking my Excel a little bit, and the FAILED profile on the first tab, looks to have been some fluke with the i1 pro 2 or with the software.
I've come to this conclusion by comparing the i1 pro 2 measurements present on each of the 3 tabs. Bottom line is:
The x,y measurements on tabs 2 and 3 (those on tab 3 were created the next day) match up very good. Even without putting the i1 pro 2 back to its exact same location.

Quick legend:
tab 1 = failed profile.
tab 2 = good profile
tab 3 = verification of tab 2 the day after

If someone is interested, I'll post the updated Excel, but I think this will not teach you guys anything new:-)

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #18 of 25 Old 03-28-2013, 12:47 PM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,431
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 78 Post(s)
Liked: 299
One example of non-linearity between profiled probes is if the colorimeter has a response of say 1.0 at light level Y1 and 0.47 at light level ( 0.5 x Y1). If the i1pro is exactly linear you would get two different matrices when profiling at those two light levels. Typical detectors have full-scale non-linearities of just a couple of percent but they aren't perfect.
zoyd is offline  
post #19 of 25 Old 03-28-2013, 01:06 PM
Advanced Member
 
realzven's Avatar
 
Join Date: Mar 2010
Posts: 519
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 84
Zoyd what is your acurate method to profile your d3 with your i1 pro with a plasma ?
realzven is offline  
post #20 of 25 Old 03-28-2013, 01:17 PM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
Hi Zoyd,

I'm not sure I understoon that properly. We can't make a non-linear colori- or spectrophotometer beter with profling as it is done now: we only measure 75% or 100% stimulus for white and the primary and secondary colors and that is it. If the meter is not linear at all then that would make it kind of useless. But you know a looooot more than I do so I'm not getting your point:)

Unless someone comes with something like a cube for meterprofiling as Lightspace CMS and Calman 3D lut things.

So incorporating Y for a meter profile is only because it is assumed the spectro is also more accurate for reading Y? Which is actually not all that important as long as the meter is linear. That is what I'm regarding as "true" at the moment.

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #21 of 25 Old 03-28-2013, 01:18 PM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,431
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 78 Post(s)
Liked: 299
I do what other folks have mentioned, roughly get the same size patch measured and use 75% stimulus. The variations I've seen at typical distances/screen areas and luminance levels are small enough that I don't obsess about it.


@Jeroen1000: I was just using that as an example of the kind of systematics that Smokey Joe was talking about and you are right, we assume those are small enough not to have to correct for them.
zoyd is offline  
post #22 of 25 Old 03-28-2013, 01:42 PM - Thread Starter
Senior Member
 
Jeroen1000's Avatar
 
Join Date: Nov 2008
Posts: 314
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 12
oh, I see thanks. But even if they were large (only talking about errors for Y here) that still would not really matter as long als the Y reading varies linear with changes in brightness. So what would be the reason to correct Y if then, seeing we can't fix a non linear colorimeter...

The only reason I can think of is a correct absolute reading

My gear: Panasonic TH-42PF11EK pro plasma display. -- Iscan Duo video processor -- i1 display 3 colorimeter -- i1 pro 2 spectrometer
Jeroen1000 is offline  
post #23 of 25 Old 03-28-2013, 10:55 PM
Advanced Member
 
realzven's Avatar
 
Join Date: Mar 2010
Posts: 519
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 84
Quote:
Originally Posted by zoyd View Post

I do what other folks have mentioned, roughly get the same size patch measured and use 75% stimulus. The variations I've seen at typical distances/screen areas and luminance levels are small enough that I don't obsess about it.


@Jeroen1000: I was just using that as an example of the kind of systematics that Smokey Joe was talking about and you are right, we assume those are small enough not to have to correct for them.
Zoyd, no problem to profile in contact mode with a plasma (measuring too green because of panel structure like Joe said) ?
realzven is offline  
post #24 of 25 Old 03-29-2013, 06:31 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,431
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 78 Post(s)
Liked: 299
Quote:
Originally Posted by realzven View Post

Zoyd, no problem to profile in contact mode with a plasma (measuring too green because of panel structure like Joe said) ?

I've not seen any supporting measurements for Joe's conjecture but in any case it would be irrelevant for profiling. If the "dark area" has any influence on the radiance measured it would be independent of distance since both "light areas" and "dark areas" will change in the same proportion to the distance. In other words the fraction of dark area/light area is constant with distance so each probe will measure the same (hypothesized) error and it will cancel out through the matrix. This doesn't mean you've removed the error but just that it is an equal component in both meters once they are cross-calibrated.
zoyd is offline  
post #25 of 25 Old 03-29-2013, 11:21 AM
Advanced Member
 
realzven's Avatar
 
Join Date: Mar 2010
Posts: 519
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 84
Thanks zoyd for the reply, greatly appreciated
realzven is offline  
Reply Display Calibration

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off