AVS Forum banner

Set me straight on colorimeters

13K views 210 replies 32 participants last post by  CalWldLif 
#1 ·
OK, I get the fundamental operating principle part and the purpose of the various types of color temperature measuring devices.

Where I'm stuck is in figuring out which one to get, especially avoiding things that are locked to a particular software and can be used on more than one computing platform. Ideally, something that with little trouble, could be used to indicate color temperature (balance) and light levels without needing an expensive software package to go with it.

All I want to do is calibrate LCD computer and HDTV displays in my home and not spend a fortune.
 
#2 ·
Greetings

What does a fortune mean?

$150 ..
$300 ...
$600
$1000
$6000

A fortune to some is nothing to others ... like asking how long a piece of string is.

Most seem to get the i1D3 device in the $250 range. Then you have to find the software to use ... which might be free or cost you some more money.

Then realize that this can be a lot like buying a professional camera ... having one doesn't make you a professional. Knowing how to run the hardware and software doesn't mean you know how to calibrate a TV properly.

Traffic engineering software does not teach its users to be traffic engineers first. Books on brain surgery do not teach you to be a doctor first.

Don't go into this with unrealistic expectations and then get mad when those expectations are not met.

Regards
 
  • Like
Reactions: tubetwister
#6 ·
If you are on a budget, and are willing to put in a bit more effort learning to use your meter with free software, you can just pick up the Colormunki Display like I did and use HCFR for calibration, and dispcalGUI + ArgyllCMS for profiling monitors. Both work fantastically well, and the Colormunki Display has the same internal hardware as the more expensive i1D3.

The caveat is that the Colormunki Display will not be supported by any 3rd party software like Calman, Chromapure, etc. However, if you are on a budget and are willing to put in a bit of effort learning to calibrate your display and use HCFR, you can get excellent results this way.
 
#14 ·
Ok by "OEM" version, does that mean it is restricted to only one software vendor?

I just hate getting locked into a piece of hardware that only works with one vendor.

Frankly, I'd be happy with something that just read out what the color temperature is with no further adieu. I'd never buy a socket set that didn't work with other brands of socket wrenches.
 
#16 ·
Frankly, I'd be happy with something that just read out what the color temperature is with no further adieu. I'd never buy a socket set that didn't work with other brands of socket wrenches.
You'd want x,y,Y data.

Color temperature is like longitude. If you only ever had longitude, you'd never be able to navigate to anywhere.

The X-Rite meter program had the OEM meter for 3rd parties to use, and the Retail meter for X-Rite software.

Many 3rd party softwares also support the retail meter at this point.
 
#18 · (Edited)
I don't get why people rely on some perfect charts achieved with colorimeters.
I myself am using the d3 and I'm very frustrated that in my country there's no one willing to rent a spectro to profile my meter.
Unless my sony br player isn't too much off (and have checked with my laptop as a pattern generator too and it's pretty much the same) I simply don't find the calibrated-perfect charts picture to be accurate for my eyes.
So, unless paired with a spectro I wouldn't rely too much on a disp pro colorimeter.
I would love to see some charts with a profiled vs edr-plasma I1disp pro on a Panasonic plasma.
 
#19 ·
The problem is that calibration with a non accurate meter can still yield you nice looking charts that don't transfer into a good looking picture.

The only combination that satisfied my picky nature was to use a i1D3 profiled off of a Jeti 1211 using a eecolor processor to hole LUTs calibrations. Additionally, I'm programming different gammas into the LUTs box to handle how different content is mastered.
 
#21 ·
So you get my frustration. .:)
I sure do. Kept doing what I was reading here on AVS and knew that my results weren't very good. Many times they were worst than out of the box with just a few minor adjustments.

Eventually I figured out that calibration software doesn't know how accurate your colorimeter is and colorimeters vary in inaccuracy. Then I found out that using an i1pro2s to profile a i1d3 don't work well on plasmas. Maybe on other technologies it works well but just not on a plasma. Something about peaky spectral light distribution.

Then you figure out how much its going to cost you to get the "right" spectrometer along with the upgrades to your software. That's about the cost of 3 high end displays or a down payment on a house.:rolleyes:

That's about the time you wish you didn't buy any of this stuff and just hired a good pro calibrator. Emphasis on "good" as not all are. Of the three calibrators I've dwelt with, I'd only consider using Chad B. Well equipped and knows his stuff.
 
#23 ·
Yes, I too would like to see the evidence that you can't profile a d3 from an i1pro on a plasma. That would essentially mean that you can't use an i1pro on a plasma, and I've never heard that was the case.

As far as I know, peaky light distribution is what makes colorimeters like the d3 untrustworthy on displays that have spectral distributions different from the one's that their built-in tables are designed for, but that's exactly why you profile the colorimeter from the spectrometer.
 
#24 ·
You can, but the question is how accurate will it be?

The size of the measurements in the i1pro and i1pro2 are 10 nanometers in width. In a reference spectro such as the jeti 1211, it's 5 nanometers in width. So the precision of the measurements are affected.

I know that between me and 2 other i1pro2 users, the i1pro2 under measures red. I also recall when Silly Sally sent his i1pro2 to Klein (makers of the K10A) Klein found that his i1pro2 under read the color red when compared to their reference grade spectro for plasmas. That translates into over correcting and having too much red in the picture. I've not been able to verify this but one i1pro user claims that it under measures green. There was a design change in the i1pro2 from the i1pro but I would have expected the error to be the same.....unless XRite was trying to correct green and got red wrong in the process. If you have an older i1pro, I recall a mod that does slightly increase accuracy. You'll have to google for it for more info.

As they say, the proof is in the pudding. Profile your colorimeter from a i1pro2 and observe flesh tones. You'll find that you don't have the range of tones/colors that you have in real life. In people of color, that shows up as a more red and less brown. In caucasians, you don't have the various shades of skin tones and there is a definite red bias. Watch the Today show where you've got a pretty good range of skin tones and errors seem to stand out better.

At one time I was trying to quantify the amount of error in red so that I might enter corrections to the i1pro2 reference readings. Nobody really seemed interesting in quantifying the amount of error so I was unsuccessful in getting this. I also tried getting a profile off of the jeti/i1pro2 and transferring it as a correction to i1pro2 to i1d3 but it failed miserably. Partly because the i1pro2 didn't profile well from the jeti 1211 and I think partly because the software didn't take the offset correctly....or maybe I got the math wrong. lol

As far as accuracy is concerned, you also have to keep in mind that traditional 6 color calibrations only gets you accurate at one lumanance level. If it turns out that other luminance levels are off, you'll likely want to go the LUTs calibration method for accuracy. The LUTS processor from eecolor cost about $700. Not all displays need this but its simple enough to measure with some software and decide on whether or not to go the LUTs route. Only problem with using a LUTs processor is that I watch a fair amount of Amazon prime and Netflix off of my Samsung's internal smart features. You can't exactly insert LUTs processing to those apps. Will probably get an external box in the future for those apps. Given the picture quality on my F8500, I don't really have any interest in upgrade displays. So what I spent to get this picture would have bought another display but then it would have needed calibration too.

Bottom line is don't expect to get much more than pretty charts trying to do it on the cheap. If you want accuracy (which is the whole point of calibration in the first place) you're going to have to spend more and very likely cost way more than what you would pay a pro calibrator for even a traditional calibration.
 
#25 ·
Greetings

When an imperfect device tells you that you are perfect ... are you? :D

My Jeti and CS 200 certainly weren't. CS 2000 certainly isn't. Don't delude yourself into thinking that perfection is attainable ... it ain't.

Regards
 
  • Like
Reactions: chunon
#33 ·
#34 ·
Let's not discourage folks from doing DIY. Maybe the i1pro isnt reference but it is not junk either. Chad used an i1pro for calibrations a couple of years ago knowing him I find it hard to believe he used something that was not fit for a plasma . Buzz says the i1pro works fine on his plasmas I believe him.
 
#36 ·
Well, now that we've pretty much thoroughly hijacked this thread, I thought I'd throw my two cents in.

I do this as a hobby so I make no money from it. I have an i1pro and a d3. I like the d3 because it's fast and has good low light sensitivity. I don't like the d3 because it's a colorimeter and 1) is not necessarily accurate for the particular spectral characteristics of a display technology if it doesn't have internal tables that match that technology, and 2) the pigments on a colorimeter's filters drift over time so it needs to be periodically recalibrated. I may be able to talk myself into spending the money for a colorimeter as a hobby but I'm not at all interested in sending it back every year for adjustment. That would be fine if I was in business but for a hobby not so much. I like the i1pro because it is a spectrometer and so does not need specific tables for individual display technologies. It also has a reputation for being accurate even after years of not being calibrated. You can send it in the Xrite and they'll send you a piece of paper telling you it's still accurate and soak you a few hundred dollars, but what they won't have to do is adjust anything. That's ideal for my concept of a hobby. The drawback of the i1pro is its lack of low level light sensitivity. It makes me cry that I can't calibrate 20% or below accurately with my i1pro.

So profiling my d3 off of my i1pro is the best of both worlds. When the d3 drifts the i1pro meter correction will adjust for it. No sending the d3 back to the shop. I also get reasonably accurate and fast readings down to 5% so I can't really ask for more. Is the i1pro the most accurate device for the job? Of course not. It didn't cost 5 times the price of my TV (or more) so I don't expect it to be. Is my picture excellent? I think so. Looks like real life to me (well, if the input material is up to snuff). So I'm happy and that's what matters.

There's a saying I think we need to keep in mind both while purchasing calibration equipment and while doing a calibration: "the perfect is the enemy of the good." Everything we work with, the measuring equipment, the display equipment, and the media being displayed is a compromise. Be happy with "good". It's better than anyone before us ever had.
Roy
 
#37 ·
Roy or someone else. Please do a grayscale measurement in 5% steps with d3 plus edr plasma (I suppose you have a plasma) and d3 profiled with the spectro. I'm very well aware that there are differences unit-to-unit but I'm just curious.
Also, maybe it is some sort of hijacking this thread but it is useful information about colorimeters in generally, pretty graphs and what some of us might think to have achieved with a budget solution. This meter is very appreciated but, myself as a user of it, for my eyes it seems to be a lot to improve in its accuracy.
 
#41 · (Edited)
The truth is out there

One of these days I am going to put a definitive end to this topic. I have profiled hundreds of D3s using a JETI 1201 on several different display devices. I have all of that data archived. When I get the time someday I will publish that data. I can say a few things now.


  • There is noticeable unit-to-unit variation. They do not all measure the same.
  • They are more accurate with some displays than with others. For example, they are quite accurate with UHP front projectors. They are less accurate with plasmas. The accuracy also varies not only with different display types--e.g. standard CCFL LCD vs. LED LCD--but it also varies between different manufacturers of the same display technology.
  • Having said this, in general they are fairly accurate without any correction. Typically, errors do not rise above xy0.006, but occasionally are as large as xy0.010. It is not uncommon with several display types to see negligible errors in the xy0.002-0.003 range. With some colors on some displays they are spot on--no error at all compared to a reference device (this is often true with blue).
  • I have no hard data to back this up right now, but my sense is that the retail meters are slightly more accurate than the OEM meters. This is counterintuitive, so when I finally get around to doing this study I'll look at this.
 
#43 ·
One of these days I am going to put a definitive end to this topic. I have profiled hundreds of D3s using a JETI 1201 on several different display devices. I have all of that data archived. When I get the time someday I will publish that data. I can say a few things now.


  • There is noticeable unit-to-unit variation. They do not all measure the same.
  • They are more accurate with some displays than with others. For example, they are quite accurate with UHP front projectors. They are less accurate with plasmas. The accuracy also varies not only with different display types--e.g. standard CCFL LCD vs. LED LCD--but it also varies between different manufacturers of the same display technology.
  • Having said this, in general they are fairly accurate without any correction. Typically, errors do not rise above xy0.006, but occasionally are as large as xy0.010. It is not uncommon with several display types to see negligible errors in the xy0.002-0.003 range. With some colors on some displays they are spot on--no error at all compared to a reference device (this is often true with blue).
  • I have no hard data to back this up right now, but my sense is that the retail meters are slightly more accurate than the OEM meters. This is counterintuitive, so when I finally get around to doing this study I'll look at this.

Would be very interesting to see that data
 
#46 ·
There are some people that simple can't have their meter profiled. So publishing such findings will be very very welcomed and useful.
What I would like to see and will put my mind at rest will be a comparison on Panasonic plasmas (maybe the most selled plasmas around the world) with the correction provided by x-rite with the plasma EDR and the profiled with a spectro matrix.
 
#59 ·
I don't know about any arguing but getting info on some things said has been like pulling teeth.
The i1pro and pro2 have been the meters the average DIYer relies on.
And the Plasma the type TV my gut says has the highest percentage of DIYer calibrators.

so it is expected that any info that brings into question the best tools available for the non pro would
generate some questions.
 
#61 · (Edited)
The i1Pro and i1Pro2 are approved by ISF and THX for professional use but if I thought my i1Pro wasn't "close enough" I'd own a 10-20k spectro. I've found it adequate on all display types except for laser.

I suspected there was a meter deficiency when I first calibrated a VT30 but it turned out to be display error. Ever since then I've used Color Girl Number 2 from the Qtec Hi-Definition Reference Disc (Blu-Ray) for one of my reference contents. Her flesh is quite pale and any Red problems show up like a sore thumb.

Color Girl Number 2

Flower Plant Botany Floristry Lip
 
#71 ·
The i1Pro and i1Pro2 are approved by ISF and THX for professional use but if I thought my i1Pro wasn't "close enough" I'd own a 10-20k spectro. I've found it adequate on all display types except for laser.
Any sort of laser display is likely to be trouble in terms of observer variability. The standard observer curves work fine for most people and normal display technology, but add critical observers to wide gamut laser displays, and a single set of standard observer curves are likely to show their limitations.
 
#62 · (Edited)
My feelings, if a profiled i1D3 (from a i1Pro, especially the Rev D) is having issues on a Plasma then it's a problem with the profiling, software or the i1D3.

There were some strange reports when the i1Pro 2 was first released, I suspect there were software issues during this time.

I'll still be recommending having a i1Pro with the i1D3 (if budget allows).
 
#63 ·
Yikes...I've been considering buying a retail i1 Display Pro to learn how to calibrate TVs as a hobby, and this thread has definitely taken the wind out of my sails. What's the point in dropping $190 for a colorimeter that itself needs to be calibrated yearly, and has seasoned folks debating on the accuracy of those calibrations to begin with. I think I'll just stick with a pro even though I think I have the patience and aptitude for calibrations. Not willing to invest so much time and money into something that might yield marginal results.
 
#65 ·
Yikes...I've been considering buying a retail i1 Display Pro to learn how to calibrate TVs as a hobby, and this thread has definitely taken the wind out of my sails. What's the point in dropping $190 for a colorimeter that itself needs to be calibrated yearly.
They don't need to be calibrated yearly. NIST certificate are only valid for a year, because you can't guarantee something longer than that. But the C6's that have come back in for recertification have all passed, except for those that are physically damaged. So we have meters that are 2 years old and still very close to how they came out of the package.

This isn't true for lesser meters like spyders, or any meter that has exposed filters, but the sealed optics in the i1 Display Pro are quite durable.

I think the other thing to note is that you've seen amateurs debating the validity of the meter's accuracy. People the people that actually have access to high end hardware on a regular basis (there is a CS-2000, 15ft away from me at the office) are the ones that will tell you the D3 is really pretty good out of the box.
 
#66 ·
@sotti, Tom and gwgill.

When using your respective software to make a meter profile of say C3/I1Pro using a Plasma display.
How much variation is there between multiple meter profiles sessions, that are done with the above same meters on different days. ?

iow, could the above lead to a larger error factor with the readings of the profiled active meter. ?

ss
 
#74 ·
When using your respective software to make a meter profile of say C3/I1Pro using a Plasma display.
How much variation is there between multiple meter profiles sessions, that are done with the above same meters on different days. ?

iow, could the above lead to a larger error factor with the readings of the profiled active meter. ?
The biggest variable I have found is warm-up time of the display. Always wait at least a half an hour before taking measurements and then do that consistently.
 
#68 ·
I don't have access to a reference spectro but I can say that I've used my i1pro (rev D) and C6 combo on three plasmas (an old 2008 Panasonic, the 2013 S60, and my current 2013 Samsung F5300). I've never seen anything visually to suggest the accuracy of my spectro or profiled colorimeter is off in any visible manner (test pattern or reference material). The fact that many pros still use the i1pro1/2 as their reference meter should speak for itself.
 
#73 ·
This is mountain out of a mole hill stuff

On average the i1Display Pro colorimeter performs well enough for 95% of users

The next 4% of DIY'ers and some Pro users crave reference grade performance so pair the colorimeter with a basic grade spectro to gain a little more accuracy.

The next 0.99% who are some DIY'ers and mostly Pro users with enough spare cash to purchase reference equipment. They gain ever decreasing circles of performance for ever increasing cost. I'd point out that pro users reason for having reference equipment is to minimise risk of wrong readings when dealing with an assortment of technologies.

The final 0.1% of users have ultra top end reference equipment to test the likes of the i1Display Pro.

Currently, the best buck for performance is the i1Display pro, period. This is the device I don't hesitate to advise to purchase for any DIY calibration situation. Sure there is error, but no measurement device is not without error, so keep this in perspective.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top