Re: LED displays: some comments and observations.
Re: LED displays: some comments and observations.
- Subject: Re: LED displays: some comments and observations.
- From: email@hidden
- Date: Fri, 07 Dec 2007 15:00:23 +0000
Hi Roger,
>
> Do you mean 'aligned physically' as in the LEDs being somehow "off axis"?
>
No, I mean spectrally. The filtration used in the current panels is not normally optimized for LED spectra. This results in some inefficiencies and some other issues. Typically the blue filter has some signal from the green led. The bandwidth of the filter is too large. This means that there is some gamut loss (although quite small). The green filter is often cutting off part of the green emission because of the wavelength shift to a lower wavelength. This cut's down on the maximum luminance.
>
> No gains in the red or blue primary?
>
There are some gains, particularly in the red, but if you look at the data in u'v' space rather than an x y representation, you would see that they are not quite so large from a visual standpoint.
>
> So, the implication for metrology is to wait, what, a good hour before
> calibrating and profiling?
>
The same could be said for critical viewing for proofing as well.
>
> But in spite of this shortcoming you were able to find an acceptable set of
> correction factors for the XL for your two filter-based colorimeters?
>
We picked the standard that the customer wanted us to use. This issue is a real metrology issue and we will be investigating it this year. I don't think that we have any good data on the visual implications and there is going to be a lot of variability in the displays themselves due to the LED sorting issues and thermal issues.
>
> You mean 'observer metamerism'?
>
I didn't want to use that term, "metamerism", because of the "religious scope" the term has taken on. The observer diffirences may not be a real problem, but we do need to look at it. The problem is that it is an extremely difficult test to run with any kind of certainty.
>
> Thereby accentuating the need to rely on the instrument as a independent
> judge.
>
Actually, what I am saying is that you can't really trust the instrumentation either. These measurements are going to require very narrow band response characterizations. This won't happen with an array based device. The basic instrument will have to be based on a dual monochrometer with a good ole' photomultiplier tube. I know that we have this stuff somewhere in the company. I'm just going to have to hijack it for a few months to convince myself it is not a problem.
Take care.
Tom L.
-------------- Original message ----------------------
From: Roger Breton <email@hidden>
> Hello Tom,
>
> Thank's for taking the time to share those valuable insights with this List.
> I, for one, always appreciate reading what you have to say about color
> management, especially displays metrology.
>
> Please allow to ask you to clear a few things, if I may.
>
> > The LED
> > technologies do not optimally align with the filter sets of the LCD
> > panels in use today.
>
> Do you mean 'aligned physically' as in the LEDs being somehow "off axis"?
>
> > ...The reason the LED panels are hard to get,
> > is that they hard to build at any price, and it is nearly impossible to
> > make money at the price they are selling at.
>
> Sounds like a tough business model.
>
> > The increase in Gamut achieved by the LED backlight is found mostly in
> > the green.
>
> No gains in the red or blue primary?
>
> > Rather than narrowing the bandwidth and increasing the
> > output, physics has demanded that the dominant wavelength of the LED
> > shift to a lower dominant wavelength than a filtered CCFL lamp.
> > (Roughly 550nm -> 520-535).
>
> Different chromaticity then.
>
> > This results in an increase of gamut, but
> > there are some real problems lurking here. The range of dominant
> > wavelengths for the LED are a strong function of temperature and
> > chemistry. For any given LED, the center wavelength and optical
> > bandwidth of the signal changes with temperature. From cold start to
> > maximum operating temperature, I have measured an xy chromaticity change
> > on the order of .013 . (Note, that's one zero beyond the decimal point,
> > not two.)
>
> So, the implication for metrology is to wait, what, a good hour before
> calibrating and profiling?
>
> > If you look at the spectrum of the various choices of green
> > leds you will see that their rising edge crosses the rising edge of the
> > y-bar CMF and falls squarely in the interaction region between the z,x
> > and y cmf's . What does this mean? It means that it is very difficult
> > to measure chromaticities accurately with any instrument at any price.
> > The difference between two very expensive spectral instruments can
> > easily be .008 when measuring the same source instantaneously.
>
> The difference between a PR-655 and a CS-1000 could easily be .008 in y
> chromaticity? On the green primary alone? I guess it amplifies whathever x,y
> errors in the instrument calibration in this region of the spectrum.
>
> But in spite of this shortcoming you were able to find an acceptable set of
> correction factors for the XL for your two filter-based colorimeters?
>
> > There is
> > another issue as well. If you believe that the cmf's are based upon
> > human visual experiments and represent an "average",
>
> We all have to believe that!
>
> > one should expect a
> > rather high degree of observer differences when looking at the same
> > displayed color.
>
> You mean 'observer metamerism'?
>
> > In short, you got your gamut, but it's probably not
> > the same gamut for the guy sitting next to you. If you have an issue
> > with inter-instrument agreement now, matters will only get worse with
> > these new technologies.
>
> Thereby accentuating the need to rely on the instrument as a independent
> judge.
>
> >
> > Tom Lianza
> > X-rite corporation.
>
> Roger Breton
>
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden