"But then those correction curves would only be strictly valid for a given pair of measurement device and display. " Exactly. Which is why I think they worked better in the days of CRTs with less variability between phosphors. These CRTs also typically had a sRGB-like response. More channels would make a colorimeter more robust, yes. More channels would essentially allow for more 'sampling' across the spectrum, possibly reducing the inaccuracies introduced by narrower primaries & unmatched filters. -Rishi On Mon, Jun 20, 2011 at 2:15 PM, Knut Inge <knutinh@xxxxxxxxx> wrote: > "Needed" in the world of commerce can be substituted with "what are we > able to sell and make money off of". My guess is that there are a lot > of colorimeter sales to people (like me) who dont really know what to > expect, who use it one or two times then put it in a box in the attic > because that cant find a way to use it that contribute to image/video > editing/vieweing in the way that the had been hoping. > > But then those correction curves would only be strictly valid for a > given pair of measurement device and display. Wide-gamut LCD displays > have narrower (?) primaries (either color filters or peaky > backlighting) to allow for more saturated colors. Introducing more > irregularity to the frequency domain would make it generally harder to > sample "color" in a XYX/perceptually meaningful way. Further, any > assumptions made in colorimeters about typical display behaviour could > be thrown off by these displays. > > I thought that the (7channel?) approach of the Spyder was supposed to > make it more robust against such display variation, although the > Hansen test seems to indicate that they have got some assembly-line > control issues. > > On Mon, Jun 20, 2011 at 10:36 PM, Rishi Sanyal <rishi.j.sanyal@xxxxxxxxx> > wrote: >> I was under the impression that correction matrices are always needed >> for any colorimeter due to the fact that the transmission profiles of >> the filters can never accurately approximate the XYZ functions of the >> human eye-brain system. This is certainly not hard at all to believe, >> as the X, Y, Z spectra are complicated, not simple gaussian curves or >> spikes at certain wavelengths. >> >> Therefore a correction matrix is needed to essentially approximate the >> total signal (# of photons, e.g.) a device with an excitation profile >> of, say, X (in other words having a transmission profile that looks >> like the excitation curve for the X function) would 'see' when given a >> certain color. >> >> Correct me if I'm mistaken. >> -Rishi >> >> On Mon, Jun 20, 2011 at 1:30 PM, Alan Goldhammer >> <agoldhammer@xxxxxxxxxxx> wrote: >>> NEC do source a colorimeter from X-Rite that has been modified with a filter >>> to work on wide-gamut displays. I've been using it with dispcal to profile >>> mine and it works fine without a correction matrix file. I have followed >>> the Ethan Hansen evaluations with great interest but even he pointed out >>> that these corrected colorimeters do a good job for the monitors they were >>> designed for. >>> >>> Of course if you have the appropriate instrumentation you can prepare your >>> own correction matrix using the tools that Graeme has developed. >>> >>> Alan >>> >>> -----Original Message----- >>> From: argyllcms-bounce@xxxxxxxxxxxxx [mailto:argyllcms-bounce@xxxxxxxxxxxxx] >>> On Behalf Of Rishi Sanyal >>> Sent: Monday, June 20, 2011 3:59 PM >>> To: argyllcms@xxxxxxxxxxxxx >>> Subject: [argyllcms] Re: Dell U2711 - is it any good? >>> >>> The correction matrix is something they download into the firmware on >>> the colorimeter. Try: >>> http://lumita.com/site_media/work/whitepapers/files/xrite-wp-3a.pdf >>> >>> On Mon, Jun 20, 2011 at 12:15 PM, Knut Inge <knutinh@xxxxxxxxx> wrote: >>>>> Knut, the Spyder 3 is a colorimeter, yes? As a colorimeter, it's not >>>>> going to work well on wide-gamut displays unless the correction matrix >>>>> supplied by the manufacturer is for wider-gamut displays. Even then, >>>>> there's bound to be some inaccuracy. Your best bet is to use a >>>>> spectrophotometer to either make the profile, or to at least make the >>>>> correction matrix for YOUR monitor + colorimeter combo, then use the >>>>> colorimeter with that correction matrix. >>> >>> >>> >> >> > >