great white balance but primaries a bit off. - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 22 Old 05-23-2013, 08:42 AM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
Been trying to calibrate my crt monitor (FW900), and here are my results. I'm curious why the green and red primaries are off even though my greyscale looks great. From what I have gathered, a good greyscale (i.e. good white point balance across the whole luminance range) is the foundation for color accuracy. Is there a reason that hitting the target primaries is somewhat independent from white point balance?

And is there anything I can do about it?

I'm using HCFR and a DTP94.









spacediver is online now  
Sponsored Links
Advertisement
 
post #2 of 22 Old 05-23-2013, 10:45 AM
AVS Special Member
 
airscapes's Avatar
 
Join Date: Dec 2008
Location: Philadelphia
Posts: 4,612
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 109
I would not put to much faith in the readings of that meter unless it was profiled to a spector. That said, those primaries are not that far off and you are only measuring 1 point. A saturation sweep would tell you a lot more about how the display really tracks.
These 2 articles will explain in detail what I am talking about.
Why you need to profile
http://www.tlvexp.ca/2012/04/do-calibration-tables-really-work-for-tri-stim-devices/
Why you need to check with Sweeps
http://www.tlvexp.ca/2013/03/color-management-system-pie-eat-half-or-all/
JimP likes this.
airscapes is online now  
post #3 of 22 Old 05-23-2013, 12:13 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
thanks for the reply. If I understand the first article correctly, profiling my device to a high grade spectrophotometer would mean making a custom calibration table that quantifies the offsets between my dtp94 and the spectro on my particular display device.

I'd then load up these tables using some function in HCFR before taking any measurements.

Is that correct?

I've read that spectros are noisy at lower luminance levels, and that the dtp94 actually excels at shadow detail. Is there any particular spectro that you'd recommend profiling against?

I'll take saturation measurements later this evening and post them here.

For what its worth, my DTP94 seems very sensitive - changing the RGB gain/bias offsets by even one button click results in visible changes in the RGB levels. Its also consistent, given that I get the same readings when taking multiple measurements.

So I understand how it may not be accurate (it may just be taking reliable measurements that are consistently ill calibrated by a certain amount). But that to me doesn't explain the discrepancy between the great RGB levels I've measured, and the results from the primary color targets. After all, shouldn't both those measures be consistent with each other (if not with accurate "real world" objective standards)?
spacediver is online now  
post #4 of 22 Old 05-23-2013, 12:48 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,605
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Quote:
Originally Posted by spacediver View Post

thanks for the reply. If I understand the first article correctly, profiling my device to a high grade spectrophotometer would mean making a custom calibration table that quantifies the offsets between my dtp94 and the spectro on my particular display device.

I'd then load up these tables using some function in HCFR before taking any measurements.

Is that correct?
Yes it is.
Quote:
Originally Posted by spacediver View Post

I've read that spectros are noisy at lower luminance levels, and that the dtp94 actually excels at shadow detail. Is there any particular spectro that you'd recommend profiling against?
The DTP94 is better than an i1 Display2, not as good as a Chroma5. Therefore it's also not as good as a i1 Display Pro, or even the new Colormunki Smile. It was a good meter when it was a new design, but now I believe that design is almost 8 years old.

For the spectro to use, the obvious choice would be an i1Pro, it's supported in HCFR can be found on ebay relatively inexpensively.
Quote:
Originally Posted by spacediver View Post

For what its worth, my DTP94 seems very sensitive - changing the RGB gain/bias offsets by even one button click results in visible changes in the RGB levels. Its also consistent, given that I get the same readings when taking multiple measurements.
The DTP94 has good sensitivity and repeatability, neither of those say anything about accuracy, but it does mean it will be reliable once profiled.
Quote:
Originally Posted by spacediver View Post

So I understand how it may not be accurate (it may just be taking reliable measurements that are consistently ill calibrated by a certain amount). But that to me doesn't explain the discrepancy between the great RGB levels I've measured, and the results from the primary color targets. After all, shouldn't both those measures be consistent with each other (if not with accurate "real world" objective standards)?

I don't understand this question.

RGB balance is based on chromaticity from the theoretical target primaries and has nothing to do with your actual primaries.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is online now  
post #5 of 22 Old 05-23-2013, 02:33 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
thanks for the detailed response, and valuable information!

With respect to my last question which didn't make sense, I think I get it now.

One can have a perfect balance of primaries which themselves do not have the correct chromaticity.

So, assuming I sort out my measuring devices, and still run into this problem (good RGB balance but not so good primaries), how would I go about calibrating my system so that I get the best of both worlds? Keep in mind that, short of going the winDAS route (equivalent to service menu of tv systems I think), the only hardware controls I have access to are RGB gain and bias offsets and brighteness/contrast.

I can imagine doing something like the following:

Display a red primary at high saturation and high luminance. Tinker with RGB gain until I get pure red on my readings.
Display a red primary at high saturation and low luminance. Tinker with RGB bias until I get pure red on my readings.

Rinse and repeat for other primaries.

But will doing this mess up the beautiful RGB balance for the greyscale I've obtained?
spacediver is online now  
post #6 of 22 Old 05-23-2013, 02:46 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,605
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Not really much you can do about primaries unless you have some sort of external control. Even then, you can only bring primaries closer to white, you can't push them back out towards the edge of the gamut. The color of primaries is a physical characteristic and you can only change them by mixing other primaries into them.

Windows doesn't provide any control over primaries, the best you can do is create an ICC profile and ICC aware apps, will do their best.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is online now  
post #7 of 22 Old 05-23-2013, 03:04 PM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 19 Post(s)
Liked: 226
Let's assume your measurements are accurate (not saying they are or aren't but if your meter is 5 years old or more, chances are the measurements aren't particularly accurate)...

The red and green are actually off quite a lot. This will be an easily visible defect with "memory colors"... colors you have seen a lot and "know" what color they should be... grass & trees, you expect to be certain shades of green. This display, if the measurements are accurate, will make all greens visibly too yellow. That display will also make reds too red. So stop sign will be a deeper red than it is in real life (as will "Target" red and "Coke" red, etc.). That error will also shift orange-ish reds towards red, so the tomato red color Chevrolet used on some of their 55-57 cars and Corvettes will look less orange than it does in real life, and more red. The errors will also oversaturate magentas and undersaturate cyans (which includes some shades of blue in the sky).

Then there's the issue of what the meter was intended to measure in the first place. If the meter was designed to be used with LCD displays, it may not be nearly as accurate when measuring a display that produces light from phosphors (CRT or plasma). I don't know anything, really, about the meter you are using. Most moderate cost meters are actually designed to be accurate with phosphor-based displays (CRT and plasma) but they don't work as well with LCD displays of any kind. With meters that measure phosphors and LCD differently, you would need at least 2 profiles or more that were generated based on a known-good meter fully capable of measuring the type of display you want to calibrate. That could mean a colorimeter optimized for each type of display technology or a spectroradiometer that is inherently accurate when measuring all types of video displays. I used a recently calibrated (a wasted $850 actually, it was remarkably accurate prior to calibration, but I have to calibrate periodically just so I can "prove it" if I have to) $14,000 meter to measure the iPad mentioned earlier. It is a hybrid meter using filters, but many more of them than are used in colorimeters and the type of filter used is not subject to much drift over time. There are so many filters used, this meter has the ability to measure any type of video display (like a spectroradiometer) accurately, but because it uses filters instead of the sort of measurements made by a spectro device, it is much better and faster at measuring low light levels than a spectroradiometer until you get into spectro meters selling for more than $25,000.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #8 of 22 Old 05-23-2013, 06:47 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
Here is the CIE diagram with saturation sweep included:



Sotti: The i1 display pro certainly looks affordable. I'm just worried about forking over that amount of cash only to find I get the same readings. I'll look into borrowing/renting one (I'm in Toronto in case anyone is game).

Doug, thanks for the thoughts. You may (or may not) have heard of the DTP94 under its other name, the Monaco Optix XR. I'm not sure what it was ideally designed for, but it can be configured through the software for CRT or LCD.

I can't tell for sure if my display matches my memory colors - I don't trust my color perception enough for that. I guess it can't be that far off otherwise it would be obvious. But perhaps I've simply adapted - I'll ask a friend to look at the skin tones and get their opinion.
spacediver is online now  
post #9 of 22 Old 05-23-2013, 07:25 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,605
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
I was referring to an i1Pro, not the i1 Display Pro.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is online now  
post #10 of 22 Old 05-23-2013, 07:59 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
ah... yes that looks quite a bit more expensive. the xrite naming system seems awful confusing btw. Is the i1pro you're referring to known as the i1 basic pro?
spacediver is online now  
post #11 of 22 Old 05-23-2013, 08:34 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,605
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Quote:
Originally Posted by spacediver View Post

ah... yes that looks quite a bit more expensive. the xrite naming system seems awful confusing btw. Is the i1pro you're referring to known as the i1 basic pro?

Basic is the software package it comes with, the meter is the same.

Check ebay for a rev D meter you can find them ~300-400 I believe.

It's a spectrometer so it's appropriate for creating profiles.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is online now  
post #12 of 22 Old 05-23-2013, 08:56 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
sotti,

Unless I'm misunderstanding the results shown here , the i1pro performs quite poorly compared to the i1 display pro, in both low and high luminance ranges:

(second graph)
spacediver is online now  
post #13 of 22 Old 05-23-2013, 09:12 PM
Member
 
cyberbeing's Avatar
 
Join Date: Apr 2006
Posts: 61
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
For what's its worth, here is the 2D and 3D the gamut of the GDM-F520 (white) compared to Rec709 (colored) as measured by an i1pro rev D. I've owed 3 GDM-F520 over the years, and they all had identical gamuts. Unfortunately these CRTs do have a gamut boundary significantly smaller than Rec709 in some areas, and there isn't really much you can do about it other than shrink the gamut even further and correct in-gamut saturations with an ICC profile or 3DLUT. Going from L* 64 to 86 (all colors) their is an increasingly larger gamut then Rec709, while L* 63 to 0 (blues, purples, and slight reds only) & L* 87 to 100 (greens only) the gamut quickly becomes significantly smaller than Rec709.
___
cyberbeing is offline  
post #14 of 22 Old 05-23-2013, 09:14 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,605
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Quote:
Originally Posted by spacediver View Post

sotti,

Unless I'm misunderstanding the results shown here , the i1pro performs quite poorly compared to the i1 display pro, in both low and high luminance ranges:

(second graph)

We've got a CS-2000 in our office, and I've never seen the i1 Pro be off by a dE of 4 on it for any display type, so I can't really put any weight into their results (Typically it's more like 1-2 at most).

Yes the i1 Pro does have a limited dynamic range, but if you were to profile a meter from it, you'd be doing it right in the sweet spot of where the i1 Pro is most effective, anywhere from 40-300cd/m^2.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is online now  
post #15 of 22 Old 05-23-2013, 10:24 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
cyberbeing: thanks for that reference information - interesting considering the f520 is considered an excellent monitor for color accuracy, and still listed among the best reference monitors by displaymate. I wonder what lagrunauer's thoughts on this limitation are. I suppose the limitation derives from the chromatic quality of the phosphors.

Joel: that looks like a gorgeous piece of hardware. I spent the better half of today linearizing the gamma of the new vpixx display in our vision lab using a much more basic minolta LS-100, and would love to experiment with something of the CS-2000 caliber.

I'm having a hard time ignoring the results from drycreekphoto. Even if they had a bad batch (they tested 10 of them), the i1 display pro still fares excellently: dE's of <2 except for low luminance wide gamut where it's 2.8.

There's a substantial discussion of this study over here. The thread starter is also the author of the study, Ethan Hansen. I'm gonna spend some time reading it, and may contact him. Based on my skimmings so far, nobody seems to have pointed out that according to his data, the i1 display pro outperforms the i1 pro in all aspects.

Assuming the data is correct, are there any other important performance parameters that would not show up on that table?
spacediver is online now  
post #16 of 22 Old 05-24-2013, 02:34 AM
Member
 
cyberbeing's Avatar
 
Join Date: Apr 2006
Posts: 61
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by spacediver View Post

cyberbeing: thanks for that reference information - interesting considering the f520 is considered an excellent monitor for color accuracy, and still listed among the best reference monitors by displaymate. I wonder what lagrunauer's thoughts on this limitation are. I suppose the limitation derives from the chromatic quality of the phosphors.

The GDM-F520 and related monitors like the GDM-C520K, have had occasional complaints over the years from users about blues tones appear more purple than normal, which makes sense considering that is exactly where the gamut is lacking the most. With the exception of the Blue phosphor, these CRTs come much closer the the gamut of Rec601 (SDTV) than they do Rec709 (HDTV). I would assume the Sony reserved the highest quality phosphors for their broadcast grade mastering monitors.
Quote:
Originally Posted by spacediver View Post

Assuming the data is correct, are there any other important performance parameters that would not show up on that table?

Colorimeters use built-in filters tuned for specific display types, which the i1 Display Pro and OEM variants complement with spectral correction tables as well. If your display matches the behavior he tables and filters were created against, the results should be very good if not excellent. The biggest limitation of such an approach, is that it will never yield perfect results on every display. If something goes wrong, you have no way of knowing without cross-checking the results against a spectro. There was a recent topic about i1pro vs i1 display pro with some links here.
cyberbeing is offline  
post #17 of 22 Old 05-24-2013, 10:11 AM
AVS Special Member
 
PlasmaPZ80U's Avatar
 
Join Date: Feb 2009
Posts: 7,135
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 62 Post(s)
Liked: 195
Quote:
Originally Posted by sotti View Post

We've got a CS-2000 in our office, and I've never seen the i1 Pro be off by a dE of 4 on it for any display type, so I can't really put any weight into their results (Typically it's more like 1-2 at most).

Yes the i1 Pro does have a limited dynamic range, but if you were to profile a meter from it, you'd be doing it right in the sweet spot of where the i1 Pro is most effective, anywhere from 40-300cd/m^2.

Over the years, roughly how many i1Pro's has SpectraCal checked? Is the number in the thousands?

If it is, that should put the Dry Creek Photo study to rest (which I have seen quoted on this forum frequently since I joined AVS in 2009)... I've also seen similar comments from Tom Huffman (of ChromaPure) regarding the i1Pro's accuracy vs. the D3 (and vs. reference grade spectros for that matter).
PlasmaPZ80U is offline  
post #18 of 22 Old 05-24-2013, 12:45 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
thanks cyberbeing, that makes sense.

PlasmaPZ, do you mean Huffman was also challenging dry creek's results? Or that he was agreeing with them?
spacediver is online now  
post #19 of 22 Old 05-24-2013, 05:50 PM
AVS Special Member
 
PlasmaPZ80U's Avatar
 
Join Date: Feb 2009
Posts: 7,135
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 62 Post(s)
Liked: 195
Quote:
Originally Posted by spacediver View Post

thanks cyberbeing, that makes sense.

PlasmaPZ, do you mean Huffman was also challenging dry creek's results? Or that he was agreeing with them?

I was saying that he doesn't consider the i1Pro to be a reference instrument nor a better choice than a D3 (or enhanced D3 to be more exact... D3 PRO) on the typical, average display. I think he did a small experiment with plasmas regarding this a while back.

However, most on this forum who have been able to compare the i1Pro to high-end spectros have found results mirroring what Sotti described.
PlasmaPZ80U is offline  
post #20 of 22 Old 06-15-2013, 08:24 PM
Advanced Member
 
S. Hiller's Avatar
 
Join Date: Jul 2003
Posts: 999
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 11
Quote:
Originally Posted by cyberbeing View Post

The GDM-F520 and related monitors like the GDM-C520K, have had occasional complaints over the years from users about blues tones appear more purple than normal, which makes sense considering that is exactly where the gamut is lacking the most. With the exception of the Blue phosphor, these CRTs come much closer the the gamut of Rec601 (SDTV) than they do Rec709 (HDTV). I would assume the Sony reserved the highest quality phosphors for their broadcast grade mastering monitors.
Colorimeters use built-in filters tuned for specific display types, which the i1 Display Pro and OEM variants complement with spectral correction tables as well. If your display matches the behavior he tables and filters were created against, the results should be very good if not excellent. The biggest limitation of such an approach, is that it will never yield perfect results on every display. If something goes wrong, you have no way of knowing without cross-checking the results against a spectro. There was a recent topic about i1pro vs i1 display pro with some links here.

On the F520, did you try it for the various picture modes and such?

Displaymate is still standing by the F520 as their reference monitor even after all of these years. FWIW, mine, surplus NOS, actually came having passed a secondary QC for photo matching work and I've never noticed any purple blues or such. Not sure what any of that means regarding the television spec, but the FW900 did make for an astonishing HDTV back in the day... smile.gif
S. Hiller is offline  
post #21 of 22 Old 06-21-2013, 10:41 PM
Advanced Member
 
S. Hiller's Avatar
 
Join Date: Jul 2003
Posts: 999
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 11
Quote:
Originally Posted by spacediver View Post

Been trying to calibrate my crt monitor (FW900), and here are my results. I'm curious why the green and red primaries are off even though my greyscale looks great. From what I have gathered, a good greyscale (i.e. good white point balance across the whole luminance range) is the foundation for color accuracy. Is there a reason that hitting the target primaries is somewhat independent from white point balance?

And is there anything I can do about it?

I'm using HCFR and a DTP94.



...

FWIW...visually compared against an apparently quite good sRGB reference (iPhone5) my GDM-F520 appears to behave as your GDM-FW900 does. Blue is visually a match, but there's a little extra green in the green, and there's a bit of orange in the red.

That said, color still looks great, e.g., a spectacular looking image of a bowl of red apples. And of course, with its truer blacks and comparatively massive dynamic range, there's a whole world of color where LCD displays cannot go...

Still such a shame they don't make these fantastic displays anymore. Hope they get the burn-in issues sorted with OLED...
S. Hiller is offline  
post #22 of 22 Old 06-21-2013, 10:59 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 786
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 66
thanks for sharing that.

re OLED, check out this thread:

http://hardforum.com/showthread.php?t=1767528

I didn't summarize the color stuff as much here, but the paper does have some meaty data on the color.
spacediver is online now  
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