Previously, my calibration algorithm was supporting only sRGB / Rec. 709 white point D65, of about 6500K.
At home I have two Dell UltraSharp 2407WFP giving me the hardest time to measure and calibrate.
Somehow, their aging CCFL backlight is throwing off every spectro or colorimeter sensor I tried and targeting a D65 white point from the measurements obtained only give a terrible blueish/pinking hue.
Added to that, some kind of dithering or backlight PWM on those displays trips off the X-Rite i1 Display Pro 3 using Argyll as a driver.
Super fast colorimeter; maybe a bit too fast in this context !
Color balance being off should not happen using the spectro measurements, as XYZ readings are not coming from a RGB-filtered light sensor but instead wavelength domain sampling.
But the spectro I have is an old EFI ES-1000 bought used for a cheap price and on this particular display the results are in contradiction with what you can perceive with your own eyes.
With this new math & code, I should be able soon to calibrate those desktop monitors I spend so much time behind.
I look forward to get here as the results of my algorithm surpass by a fair margin in accuracy every other I tried so far.