Frédéric Mantegazza wrote:
As far as I understand, the reference file is in fact the primaries of a given beamer. I think they are hoping that primaries does not vary too much from a beamer to another, so they calibrate the HCFR colorimeter that way.
It sounded something like this, which is why I didn't follow in that direction. I tend to assume that the white of any given uncalibrated display device could vary considerably, and in any case, the white does not provide enough information to setup a full 3x3 matrix.
If it is the case, I agree it is not very smart, as there *are* differences from beamer to other, and the screen also has to be taken into account... So, Graeme's method is better. I'll maybe suggest them to use it!
It comes down to how consistent the sensors are, and how representative the displays used in the calibration are in terms of spectral distribution, and how good the reference instrument is. The commercial colorimeter vendors typically spend a lot of energy wrestling with things like this. Graeme Gill.