Pages: [1]   Bottom of Page
Author Topic: observer metamerism and wide gamut displays  (Read 3828 times)
Sr. Member
Offline Offline

Posts: 299

« on: July 09, 2010, 08:58:00 PM »

What are the thoughts on this?

Does it really matter much at all?

Considering differences in probes and screen uniformity, etc. maybe it doesn't matter nearly as much as getting the expanded gamut, even if it does matter.

Do the stronger, more peaked primaries in a wide gamut make it harder to maintain similar apparent white balance between different displays both for individual and multiple viewers?
Is this the one negative with 3 primary wide gamut monitors? (apparently particular types of >3 primary wide gamut displays can be designed to reduce viewer metamerism though)

Do people usually use Eizo's 10 degree settings and NEC's metamerism setting as general practice?

Apparently it is very tricky to even design a universal metamerism function since the required function varies from person to person.

One little interesting paper I found:

I've also located a number of comments where they say that if a bunch get together and look at some calibrate sRGB and some calibrated wide gamut displays that many more people disagree on various color matches on the wide gamuts since the more intensely spiked spectrum of the wide gamut monitor makes person to person different far more apparent (people have cones that react in slightly different ways and corneas with slightly different transfer functions and sRGB monitors don't make these differences from person to person as apparent).

(COmparing my DTP94 calibrated sRGB monitor to my wide gamut one using both factory sRGB or puck calibrated sRGB mode, the white balance and certain shades certainly don't look the same at all, even if I use the metamerism switch. Of course, in this case, the DTP94 could well account for almost all of the difference. I really need an EyeOne Pro spectrophotometer to adjust both. I will do a quick calib of the other monitor with the NEC puck and then compare.)
« Last Edit: July 09, 2010, 09:56:21 PM by LarryBaum » Logged
Ernst Dinkla
Sr. Member
Offline Offline

Posts: 3140

« Reply #1 on: July 10, 2010, 08:02:32 AM »

Interesting comments and documents.

You could add retinal and functional Tetrachromacy, half the women have the first and a lower percentage actually use that extended sense.

You could add wide gamut in prints from dye based, pigment based, CcMmYK to N-color printers to the discussion too. At least for the last there has been research done to reduce metameric failure. Whether the "average" observer still is a functional quantity to measure against is another question.

Despite refining the algorithms for perceptual uniformity in time, a DeltaE number isn't so absolute:

Related to this. Some years ago I asked the question whether the colorimeter hardware (filters/LEDs) then existing for creating RGB-device printer profiles could cope with the extra 2-3 hues in N-color printers. In practice the Epson R1800, R 800, some Canon models <> Spider colorimeters. The answer then from the Datacolor representative was that his knowledge didn't stretch to the hardware (or something alike, the dpreview archive doesn't deliver the messages anymore). It isn't about wide gamut but whether LEDs/sensors/matrices created for CMYK-CcMmYK (dye) inks function on CcMmYKkkRGB (pigment) inks.

I don't think spectrometers are that absolute either, a lot of different properties per model and deviations per device as used in practice.

met vriendelijke groeten, Ernst Dinkla

spectral plots of +100 inkjet papers:

Sr. Member
Offline Offline

Posts: 299

« Reply #2 on: July 10, 2010, 03:01:51 PM »

Some interesting findings:

Comparing either native factory preset modes or NEC puck+SV II calibrations of the NEC PA241W and the metamerism switch on off to a Samsung 244T (standard gamut) calibrated with DTP94b+CEDP the two displays didn't match. Large solid patches of various grays looked pretty different and even in some photos you could see a difference an evening lit orange-brown cliffside had a reddish cast on one and a more yellow-orange cast on the other.

So then I loaded up multiprofiler and loaded the SVII profile and then turned on the metamerism switch (SVII doesn't have a toggle to turn on metamerism so you need to use multiprofiler also) and then things started to look a lot closer to how the Samsung sRGB monitor looked, although they were still a bit off. The SVII+NEC puck version looked a bit closer to the Samsung than the factory preset sRGB mode in terms of the gray scale tint.

So it seems like SVII+NEC puck may well improve upon factory preset white point and that metamerism toggle should probably be always left on (multiview does ship with metamerism toggled on as a default it should be noted too). So it seems like the eye really does not see a wide gamut display the same way as a standard gamut one and that the differences can be quite noticeable, definitely enough to bothersome without applying at least some correction. This makes me wonder about all those somewhat lower end Dell and HP that as far as I know don't have a metarmerism or 10 degree calibration choice. Might stuff editing on them not match when viewed on other displays? Only Eizo and NEC support that (NEC calls it metamerism and Eizo 10 degree view calibration) right? Not even stuff like iColor or CEDP??? Or maybe they do? If they don't then doesn't it get weird when you edit on say an HP 2475 or Dell U2410 and then view on an sRGB display or a NEC or Eizo with correction turned on??

BUT of course I used different software and different puck on each, so that could explain as much or more than metamerism.

So first I went back and calibrated the Samsung using NEC puck+CEDP this time (on NEC's site they claim their special wide gamut i1D2 works fine with third party software if used on sRGB monitors only). The result was pretty bad looking. The Samsung suddenly had a noticeable red cast. The Samsung+DTP+CEDP looked MUCH closer to the NEC+SVII+NEC than the Samsung+NEC+CEDP did!
I know on another thread here it had been said that an off the shelf i1D2 tended to under read the amount of blue present and I also see someone else reporting here today that an off the shelf i1D2 made both his wide and standard gamut monitors look red.

It seems like the NEC puck is NOT calibrated to some perfect sRGB response standard but perhaps to some average i1D2 response (maybe so those with off the shelf copies can at least give them a shot with SVII), either that or so much for the claims (not that I care so much) that the NEC version will work fine using third party software if used on standard gamut displkays. And it seems like the i1D2, in general, may well produce rather red calibrations? At least more recent copies?

When used with SVII on a NEC PA I don't notice any read cast whatsoever from the NEC puck though.

Anyway, so far, NEC PA+SVII+NEC puck+metamerism toggle and Samsung 244t sRGB+DTP94b+CEDP produces the closest match. It's still not quite the same looking though. Same thing when I compared that NEC PA combo with a samsung hdtv that has been calibrated with a DTP94b too. Of course how much of this is due to the metamerism not being perfectly corrected for and how much has to do with the fact that different software and pucks were used on both is hard to say, that certainyl could account for it all, perhaps if a spectrophotometer was used on both having metamerism toggle off would even be closer (but probably not).

I hope to get a chance to give the samsung (and perhaops nec) a quick calibration with an i1Pro, would be interesting to see what happens then.

For now I would tentatively suggest using a puck to calibrate the NEC PA with SVII rather than factory preset and NOT using the NEC puck to calibrate any displays and to keeping the metamerism toggle on as the best bet. (EDIT: see post below!)

I will see what happens if I get to try an i1Pro. That may every well change the findings, since there have been some claims that the DTP itself, while having good inter agreement, have a bit of a white point bias of their own.
« Last Edit: July 10, 2010, 06:46:18 PM by LarryBaum » Logged
Sr. Member
Offline Offline

Posts: 299

« Reply #3 on: July 10, 2010, 06:37:22 PM »


ok this time I decided to simply bring out my color checker chart and compare to the my NEC PA241Q and Samsung HDTV.

Ideally I would've viewed the CCC under a special D65 light, but I don't have one at the moment so I tried to make do with indirectish sunlight bouncing into a room.
I couldn't do this last night as all I had were fluorescent and tungsten.

Well I have to say the NEC PA241W with metamerism OFF was a closer match to watch I saw on the color checker chart than with it ON.

Although metamerism ON does make the NEC look closer to the HDTV when looking at both at once.

I can also say that both my sRGB calibration using the NEC puck and the native sRGB mode, both with metamerism off, actually look closer to the CCC than did the (non-wide gamut) HDTV I calibrated using a DTP94b puck and nothing but the monitor controls (no profiling yet).

I didn't get to compare with my Samsung 244T (non-wide gamut) as calibrated with a DTP94b and CEDP yet.

So now I'm thinking maybe it's actually better to leave metamerism OFF unless you are directly working next to a non-wide gamut monitor and staring into both at the same time....


Anyway it seems like the wide gamut NEC is perhaps actually matching closer to what my eye sees in the real world than my DTP94b calibrated sRGB displays....
not that even it really matches what i see on the color checker if I bring the chart in close, but it definitely does look like a better overall match than the DTP94 calibrated HDTV or Samsung 244T+DTP94+CEDP.

I guess I really need to compare when using a D65 lamp to really be sure, although it seems like ones with smooth spectral characteristics cost almost as much as the monitor! So I think I will wait for an overcast day at noon hah.
« Last Edit: July 11, 2010, 02:36:03 PM by LarryBaum » Logged
Pages: [1]   Top of Page
Jump to: