Help understanding calibrators

[icor1031]

My understanding:

There are tri-stim, and spectrophotometers. The latter is very expensive.
http://www.tomshardware.com/reviews/spyder4-monitor-calibration-image-quality,3581-2.html

Tri-stim are usually inaccurate, as the above link says as well as these:
http://www.tftcentral.co.uk/reviews/content/spyder3elite.htm#conclusion
http://www.tftcentral.co.uk/reviews/content/spyder3elite.htm#success

Spectrophotometers are very accurate,
http://www.tftcentral.co.uk/reviews/content/i1_pro.htm#conclusion
http://www.tftcentral.co.uk/reviews/content/i1_pro.htm#lacie



I assume because of the price that these are not spectros. Am I wrong? Why are they so cheap? Are they likely to still be good, despite their age?
Why are their calibrations so accurate, if they're not spectro? Is it an error?
http://www.tftcentral.co.uk/reviews/content/lacie_blue-eye_2.htm#results
http://www.tftcentral.co.uk/reviews/content/eye_one_display2.htm#success


Thanks!!!

 

[icor1031]

ALSO, I saw that some have a "UV-Cut Filter." What does this do? (Do I want that?)

 

[spacediver]

spectros are generally more accurate than tri stims because of the way they work. See this excellent white paper from x-rite.

You can get good second hand i1 pros for under $400 on the bay.

Good tristims can be as accurate as a spectro if it has been profiled to the display type that you are using.

One prob with consumer level spectros is that they have problems reading accurately in the lower luminances, so many people get a spectro and a colorimeter, and use the spectro to profile the colorimeter, and use the colorimeter for actual calibration.

In some cases, e.g. working on trinitron CRTs, a colorimeter like the DTP-94 comes calibrated for those particular phosphors so you don't need a spectro.

an i1 display pro is a great colorimeter and should provide accurate readings for most displays, although I would spend some time reading through the avs display calibration forum to seek more information.

As for the UV cut filter issue, see this thread

 

[icor1031]

Spacediver!
Thanks for the great reply. I read that whitepaper, but I'm still confused. I have a few questions:

I'm calibrating LCDs.



I read that spectroradiometers need to be re-certified after time, to see if they're still accurate.
Is it likely to be off by much, if I buy an old unit?

Why did the colorimeter in the tft-central link perform so horribly? Did they do it wrong? Because if not, why should I think a cheap one from ebay for $40, or even a new one on amazon for $200, will work well? Is it just that the specific model they used, is terrible?
I read the Amazon reviews about tri-stims, and people say that they get a noticeably different calibration each time, on the same monitor. Did they just do it wrong?

I read that the filters in the colorimeter can change with time. For that reason, is it a bad idea to get an old colorimeter from ebay for $40?

How do you profile the colorimeter to the spectro?



I love the concept of the colorimeter after reading the paper you showed me. I just need to understand it, to make sure I'm not wasting money.

You rock.

 

[spacediver]

Why are their calibrations so accurate, if they're not spectro? Is it an error?
http://www.tftcentral.co.uk/reviews/content/lacie_blue-eye_2.htm#results
http://www.tftcentral.co.uk/reviews/content/eye_one_display2.htm#success

btw, what makes you think these calibrations are accurate? Those delta E values are relative to a particular standard (sRGB). The data say absolutely nothing about the accuracy of the instrument itself.

 

[icor1031]

btw, what makes you think these calibrations are accurate? Those delta E values are relative to a particular standard (sRGB). The data say absolutely nothing about the accuracy of the instrument itself.

Ah..
In that case, the tomshardware site could have used a different standard, which could explain why the Delta E is so high?

 

[spacediver]

Spacediver!


I read that spectroradiometers need to be re-certified after time, to see if they're still accurate.
Is it likely to be off by much, if I buy an old unit?

hard to say - if ur a pro and charging money, get it recertified regularly. If it's for your own use, you probably won't need to get it recertified for a few years if that.

Why did the colorimeter in the tft-central link perform so horribly? Did they do it wrong? Because if not, why should I think a cheap one from ebay for $40, or even a new one on amazon for $200, will work well? Is it just that the specific model they used, is terrible?

You gave 6 TFT links. Which one? And what made u think it performed poorly?

I read the Amazon reviews about tri-stims, and people say that they get a noticeably different calibration each time, on the same monitor. Did they just do it wrong?

Probably. Important to measure the same spot on the screen each time, and wait till display has warmed up.

I read that the filters in the colorimeter can change with time. For that reason, is it a bad idea to get an old colorimeter from ebay for $40?

Depends on the colorimeter. The i1 display pro is a quality instrument and will serve u well if u treat it right.

How do you profile the colorimeter to the spectro?

Depends on the software. The basic idea is to take a series of measurements with the target instrument (the colorimeter) and the reference instrument (the spectro). Then, the software creates offset tables (or sometimes more sophisticated calculations) that are applied to future measurements on the target instrument. There are some threads about this on the avs forums.

 

[spacediver]

Ah..
In that case, the tomshardware site could have used a different standard, which could explain why the Delta E is so high?

No. Those Delta Es show how much the display deviates from that particular standard. You can have the world's most accurate instrument, and if your display isn't capable of being adjusted to meet that standard, you'll have high Delta Es.

 

[icor1031]

hard to say - if ur a pro and charging money, get it recertified regularly. If it's for your own use, you probably won't need to get it recertified for a few years if that.



You gave 6 TFT links. Which one? And what made u think it performed poorly?


Probably. Important to measure the same spot on the screen each time, and wait till display has warmed up.



Depends on the colorimeter. The i1 display pro is a quality instrument and will serve u well if u treat it right.



Depends on the software. The basic idea is to take a series of measurements with the target instrument (the colorimeter) and the reference instrument (the spectro). Then, the software creates offset tables (or sometimes more sophisticated calculations) that are applied to future measurements on the target instrument. There are some threads about this on the avs forums.

Concerning the spectro+colorimeter combination, I guess that's not a great idea if I buy a used spectro that's already a few years old, because I'll just be adding error.
So then, I think we can stop talking about spectros.


Would either of these probably still be good/accurate? (concerning filter change with time.) You mentioned that it depends on the unit.

http://www.ebay.com/itm/X-Rite-Eye-...289?pt=LH_DefaultDomain_0&hash=item3a8ed756c1

http://www.ebay.com/itm/LACIE-Blue-...883?pt=LH_DefaultDomain_0&hash=item3f28bd6a9b


Thanks, mate!

 

[icor1031]

No. Those Delta Es show how much the display deviates from that particular standard. You can have the world's most accurate instrument, and if your display isn't capable of being adjusted to meet that standard, you'll have high Delta Es.

Perfect, this answers the question that created your response about 6 TFT links. So, I removed my reply to that in the post above.


Again, I appreciate your help very much.

 

[spacediver]

Concerning the spectro+colorimeter combination, I guess that's not a great idea if I buy a used spectro that's already a few years old, because I'll just be adding error.
So then, I think we can stop talking about spectros.


Would either of these probably still be good/accurate? (concerning filter change with time.) You mentioned that it depends on the unit.

http://www.ebay.com/itm/X-Rite-Eye-...289?pt=LH_DefaultDomain_0&hash=item3a8ed756c1

http://www.ebay.com/itm/LACIE-Blue-...883?pt=LH_DefaultDomain_0&hash=item3f28bd6a9b


Thanks, mate!

best to post this question on the avs forums, and provide information about which display(s) you'll be working with. There are ppl there with a lot more experience with those instruments than myself!

If it turns out that your particular displays will work well with a factory calibrated i1 display pro, then you may be golden without a spectro.

and glad to help

 

[icor1031]

Space, you might find this useful:
http://www.luminous-landscape.com/f...7f2daac60c3b56d5cffdf63d8c8ac03&topic=53825.0

It might explain why the spyder3 measured so badly @ tft-central.

The i1display pro came out nicely, though.

 

[spacediver]

those measurements show repeatability and accuracy relative to a reference instrument.

As I've said, the TFT results tell you nothing about the quality of an instrument.

You can have a shitty instrument and get a delta E of 0 with respect to a standard like sRGB. It just means you can't trust those delta Es.

If you're planning on learning how to calibrate, you should probably read up on some basic colorimetry and encoding standards so you have a better grasp.

In a nutshell, these instruments measure chromaticity (and luminance), and are used to guide the tweaking of a display's color controls. For HD material (e.g. Blu-Ray) and web content, the target primaries (which define the points of the triangle you see in those links) are defined in advance. The goal is to tweak your display so that your display's primaries match up with those coordinates specified in the standard. Additionally, the white point is defined also (towards center of triangle).

The delta Es you see in the TFT results show how well the calibrated display matches up with those pre-defined coordinates (from the standard), as measured by your instrument. If you have a shitty instrument, you can tweak your display such that the instrument reports a perfect match with the standard, but because the instrument cannot be trusted, it means your display does not meet that standard. So you can never use those delta Es to determine the accuracy of an instrument.

I think part of the confusion is a lack of understanding of what Delta E means. It simply indicates how far away you are from a given standard. That standard could be an encoding standard (such as the triangle you see in those TFT links), or it could be a reference instrument. Or it could be a measure of how much variance there is within a given instrument's reading. You have to appreciate the context of data to understand what those figures refer to.

And yes, the spyder 3 has earned a poor reputation.

 

[icor1031]

For anyone interested, here's my AVS thread:

http://www.avsforum.com/t/1528143/x-rite-and-spyder-questions

 

[Bluesun311]

For what it's worth, there was actually some DeltaE info on the Korean review of your monitor that I linked in the thread you made asking about Contrast Ratio for it. I didn't feel like trying to explain about the different DeltaE standards that have come about over the years and I was suprised you were making entire new threads for these questions instead of just one thread with "Questions about my Yamakasi Q270 and calibration"

What the review did was give a DeltaE for whitepoint at 6500Kelvin and it was quite good.

These LG AH-IPS panels are extremely high quality, in general. Where things get fuzzy is around the edges, of course. Some of them are perfect, but not many. Almost all end up with glow for whatever reason and lightbleed because of their housing. Even the ones that get binned A+ get returned a lot.

 

[icor1031]

For what it's worth, there was actually some DeltaE info on the Korean review of your monitor that I linked in the thread you made asking about Contrast Ratio for it. I didn't feel like trying to explain about the different DeltaE standards that have come about over the years and I was suprised you were making entire new threads for these questions instead of just one thread with "Questions about my Yamakasi Q270 and calibration"

What the review did was give a DeltaE for whitepoint at 6500Kelvin and it was quite good.

These LG AH-IPS panels are extremely high quality, in general. Where things get fuzzy is around the edges, of course. Some of them are perfect, but not many. Almost all end up with glow for whatever reason and lightbleed because of their housing. Even the ones that get binned A+ get returned a lot.

I made the new thread, because my old thread had "contrast ratio" in the title. And at the time, I wasn't going to ask about the color
You are right however, I should have just edited the title and bumped that one.


I see the distance it's rated from sRGB, in your previous link. Thanks!
I have no idea if it's default, or after calibration - though. I might put it in a translator.

 

[Bluesun311]

I made the new thread, because my old thread had "contrast ratio" in the title. And at the time, I wasn't going to ask about the color
You are right however, I should have just edited the title and bumped that one.


I see the distance it's rated from sRGB, in your previous link. Thanks!
I have no idea if it's default, or after calibration - though. I might put it in a translator.

Fair enough

Hopefully, the DeltaE post-calibration for whatever you are measuring the difference between the standard you want and the actual output is close to 1 (or imperceptible)

In the color temperature section of the KoreanQ270 review (which you can have google translate and it comes out really funny) the white point difference is unconverted in terms of adhering to DeltaE calculations, but the idea is the same. The chart is, I should think, default color settings or it would be a review of the calibrator.

 

[icor1031]

The chart is, I should think, default color settings or it would be a review of the calibrator.

Because it should say near-0, even if it's not - because, the calibrator *thinks* it's correct... yeah?
And then they'd have to prove the calibrator isn't wrong.
Still, it would introduce error in either case - but less so significantly so, to merely measure it with a calibrator. I ask, to make sure I understand you. I think Spacediver mentioned this, and I didn't catch on at the time. LOL

I'm just used to tft-central reviews, which I assume test against something that's known to be correct.

 

[Bluesun311]

Because it should say near-0, even if it's not - because, the calibrator *thinks* it's correct... yeah?
And then they'd have to prove the calibrator isn't wrong.
Still, it would introduce error in either case - but less so significantly so, to merely measure it with a calibrator. I ask, to make sure I understand you. I think Spacediver mentioned this, and I didn't catch on at the time. LOL

I'm just used to tft-central reviews, which I assume test against something that's known to be correct.

I think you've got it. The calibrator knows what it's looking for and if you have calibrated your display already, then that's what it will find and your delta E will be close to 0.

Delta E is really just used to see if a tolerance level has been met. Ideally you want your color curve perfect right out of the box (like these Catleaps/Crossovers/Qnixs will generally have) not so you can be lazy, but because it will translate to a higher contrast ratio. If you start having to lower color settings to help the calibration process along, then your CR (the difference between brightest and darkest possible gradient) will suffer.

I'm glad you like your i1 display pro. I am super pleased with mine. And don't get me wrong, I'm learning as I go along here just like you--I just have a tiny head start is all. The i1 display pro has its limitations, and they are above my head and I prefer to keep it that way, but for my needs, just to make sure my tolerance levels have in fact been met in terms of ballpark figures, it's fantastic.

Now, if I were a serious graphic designer who had a lot of experience and really knew my shit and was working in Adobe RGB colorspace (why? again, above my head, but it's mostly for printing I think, because we can't display it properly without the right setup) and was constantly re-calibrating throughout the day due to changes in luminance that were outside of my control, etc. Or if I were doing any kind of optronics reviewing I would want an x-rite i1 pro 2. They are 2grand though and the i1 display pro is 1/10th the cost. Did it give me different RGB gain settings from the TFT standards offered for my display? Yep. Did it make me a custom icc profile? Yep. Did it look better to me than the one TFTcentral posted for my display? Yep. Mission accomplished!

By the way! What brightness and contrast settings did you end up with on your Yamakasi? Do you like your monitor?

 

[icor1031]

I don't have any calibrator, yet. But I think I'm going to get the i1display pro.

Thanks for the info, mate.
Do you use a DSLR (camera type), by chance?

BTW, I was unlucky. My catleap was Far off, by default.

This is the closest I've been able to get it to my U2311H or 2333T:

 

[Bluesun311]

Oh I see, shit I realize now the Yamakasi doesn't have gain settings, and it's kinda hard to quantify your brightness setting when there is no OSD

It does have brightness controls, however, and you may find that a colorimeter will simply want you to turn the brightness down instead of cutting the gains so drastically like that. You may also find that the color temperature on your other displays is incorrect, though I have no idea if that's the case.

That's pretty much the perfect use for a cheap calibration tool, though. Even if you just rented one for a couple days--you can set up different custom icc profiles for all 3 displays and then use Color Sustainer v1.04 to manage your profiles instead of allowing windows to do it. Windows does weird things with profiles and cant remember them half the time anyway because programs (particularly games) push windows color management around. So those color adjustments youve made won't even be applied in these cases. Color sustainer and a quick and dirty calibration of the 3 different displays will help a good deal, I hope.

 

[spacediver]

just a wee quibble: probably best to refer to the instruments as colorimeters or spectroradiometers, rather than calibrators. Calibration is a process that those instruments can aid in.

As for Adobe RGB, yep you'd need a wider gamut display, since the green primary is way more saturated than sRGB's standard, which is what most displays are aimed at. And it would be good for photographers/designers who want to print their work (or show case it on wide gamut displays).

 

[Sailor_Moon]

Tri-stim are usually inaccurate, as the above link says as well as these:

http://hardforum.com/showthread.php?p=1040697852&highlight=#post1040697852

ALSO, I saw that some have a "UV-Cut Filter." What does this do? (Do I want that?)

Can be helpful when measuring samples on paper with optical brighteners. With the i1 Pro 2 you are more flexible because it supports - in contrast to the i1 Pro with and without UV-Cut - different measuring conditions (=> a dedicated UV-Cut version no longer exists). However: For measuring luminous colors emitted by a display all this isn't important for you.

In that case, the tomshardware site could have used a different standard, which could explain why the Delta E is so high?

A general advise: Those graphs - regardless of a special source - can easily be misinterpreted. You must exactly know what was measured and how it was measured. A profile validation after calibration? Measurements against other RGB targets (for example common working color spaces) with or without color space transformations through the display or a CMM? Special tests for soft-proofing capabilities? Very often color distances of different measurements are placed in a wrong context.

Windows does weird things with profiles and cant remember them half the time anyway because programs (particularly games) push windows color management around

Windows can anyway only load the vcgt linearization data from the profile into the videocard LUT. Even if you manage to keep it persistent color space transformations will only be carried out by color aware software.

 

[Bluesun311]

http://hardforum.com/showthread.php?p=1040697852&highlight=#post1040697852Windows can anyway only load the vcgt linearization data from the profile into the videocard LUT. Even if you manage to keep it persistent color space transformations will only be carried out by color aware software.

Darn. I can't help wondering about an open source emulator engine like PCSX2 (Playstation 2 software) because recent video plugins for this have options for color saturation/gamma/brightness. They are very "heavy", as in, 1% or 2% gamma is very noticeable. Does this mean the software is "color aware"? Is there any way to know?

I must admit, I never concerned myself with issues of color accuracy in the past, but as my interest in actual gameplay decreases and my interest in the technology itself increases, I find that making calibration adjustments is very satisfying and noticeable, to the point that actual gameplay/content viewing/simple browsing is elevated considerably.

Do you use a DSLR (camera type), by chance

I briefly had a Canon EOS 6D and I loved it. I will have one again or something better. There's nothing like having perfect pictures of your perfect {insert thing here}.

 

[icor1031]

I briefly had a Canon EOS 6D and I loved it. I will have one again or something better. There's nothing like having perfect pictures of your perfect {insert thing here}.

If you need advice on lenses, I can probably help you.