[argyllcms] Re: Huey used with Samsung XL20?
- From: Gerhard Fuernkranz <nospam456@xxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sat, 11 Apr 2009 22:42:18 +0200
Pascal de Bruijn wrote:
> As far as I know, a colorimeter, doesn't know anything about matrix
> coefficients, it just measured colors...
This may be certainly true for $$$$ or $$$$$ (or even more) lab-grade
instruments, but the filters of the various low-cost monitor
colorimeters available on the market do not approximate the CIE standard
observer sensitivities well enough. So don't expect that these low-cost
gadgets are really accurate general purpose colorimeters. They will give
you different readings when you measure two light sources with metameric
spectra, although both measured light sources _do_ have the same color
(eventually you get, what you pay for). However, this doesn't mean that
these instruments were garbage and completely unusable. If you calibrate
them for measuring just one particular monitor model, then you still can
achieve pretty accurate readings from different displays of this
particular model. But with this calibration, the instrument won't give
you accurate readings on different monitor models. So each instrument
would need a bunch of correction matrixes, one for each monitor model to
be measured - and supplying all these matrixes would not be practical.
As a trade-off (at the cost of reduced accuracy), these instruments
usually store only two correction matrixes in their eeprom, one for CRT
and one for LCD displays (assuming that most usual CRT phosphors have
similar spectra, and that most LCD displays (with CCFL backlight) have
similar spectra too). I guess, in the future, when displays with LED
backlight become more and more popular, they'll store a 3rd matrix too -
for LCD monitors with LED backlight.
> So they probably adjusted the software, not the device...
As I already mentioned in an earlier posting, the readings displayed by
NCE and the readings from Argyll (both using the same instrument - in my
case the i1 Display bundled with my XL20) do agree pretty closely, which
would not be the case if NCE would apply additional software adjustments
to the readings it got from the instrument driver.
>
> I don't see why it would be disappointing... When you buy things in a bundle,
> does one of them _have_ to be "special" :)
In this case they should preferably fit with each other, since only a
special calibration of the instrument for measuring this particular kind
of monitor will enable accurate readings, while the standard LCD
calibration won't (note that the spectra emitted by the XL20 are pretty
different from the spectra emitted by other CRT or LCD displays, due to
the LED backlight).
Regards,
Gerhard
Other related posts:
