CCFT versus LED [was: Spectros better than colorimeters
CCFT versus LED [was: Spectros better than colorimeters
- Subject: CCFT versus LED [was: Spectros better than colorimeters
- From: email@hidden
- Date: Sun, 26 Oct 2008 20:37:23 +0000
> Well, lucky us.
>
> What I notice is that instrument designers seem to shun this forum,
> most of the time. Which is a pity because I'm stuck in interface hell
> at the moment and sure could use a helping hand.
>
> Edmund
With all due respect, I am an instrument designer and I don't shun this forum. The problem I have with the forum is that some folks don't seem to be inclined to read it and learn from it. I have stated many times on this forum that the problem with instrument measurement of LED backlights lies predominately with the green primary. The LEDs have rather broad spectra, they are not "spike like" . The gamut expansion of the LED back light comes from the fact that the dominant wavelength of the green LED has been pushed between 520 and 530 nm. This results in a number of interesting problems relating to inter-instrument agreement. The LED also shifts wavelength with temperature, so a measurement of the diode itself is highly variable. A big problem with LED back light uniformity often is the result of an uneven temperature gradient within the backlight system.
When measuring the chromaticity of a green LED, the bandpass of the spectral instrument is very critical. The problem is that the low wavelength side of the green output is very near the low wavelength rise of the y color matching function. This makes determination of the precise chromaticity very difficult. Small errors in wavelength calibration yield very large differences in chromaticity.
I have seen differences between a photoresearch pr650 and a Minolta CS 1000 greater than 0.01 with respect to an xy measurement of LED chromaticity. So you can see that two high grade spectral instruments have real issues with this technology. Our i1Pro and Color Munki instruments will generally measure closer to a PR650 than the Minolta CS1000 and that is simply due to bandwidth.
The design of end user instruments, today, is limited by effective standard instruments. The technology that can be used to make a research grade device is available, but it is neither compact nor fast. NIST, the US standards organization has built a device for such characterizations has outlined the design and anyone may freely use the design. In order to accurately measure an LED device, even with the NIST spectro, one would have to control the junction temperature of the LED. This is not a trivial undertaking.
There is no magic bullet. All measurement instruments are always behind the latest advances in display technology because the display technology changes so quickly. As some who as designed light measurement equipment my entire technical life, I can safely say that the problems are somewhat understood, it is the unique solution that is hard to find. In the end, we have to build instruments that correlate well with a research grade instrument as applied to a unique measurement scenario.( i.e Wide gamut LCD + i1Display) . Getting two spectral devices to agree is a far greater issue than calibrating a colorimeter. This is a much bigger problem with emissive measurements than reflective measurements because of the basic nature of the spectra. It is a certainly an issue we are looking at with some care, but the physical solution is not so simple....
So there you have it: yet another opinion....
Regards,
Tom L.
_______________________________________________
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