[argyllcms] Re: [argyllcms] Re: Monitor screen color unevenness/irregularity
- From: Alexey Blinov <abvgdee@xxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 17 Apr 2012 16:54:19 +0400
As Graeme suggested, here are my results in L*a*b* space (i.e., the common
Delta E), as diameters of the corresponding balls (enclosing many
measurements..):
CRT monitor spatial non-uniformity: 6, LCD: 8
Colorimeter accuracy: 0.2
Distance between the above two whites (between centers of the balls): 0.25
I now see what keywords describe this issue: "Color Uniformity Compensation".
And EIZO's white paper for their ColorEdge CG221 (see google) has the plots I
was looking for: before compensation (Delta E around 5, except at the edges)
and after (less than 1).
For the Janos' NEC PA231W, there's dramatic improvement both in color temp and
Y, like 10-15 times. Everything smooth is linear in small regions, so errors
(also Delta E) are reduced proportionally. Similar as EIZO.
So looks like calibrating/profiling a regular monitor only makes sense for some
small area, e.g. at the center. And its size is to be determined by the user
experimentally. Then scale your photos to fit interesting parts into that area.
Simple :) BTW, such experiment should ideally be done for several graylevels..
Mon, 16 Apr 2012 20:08:19 +0200 от János, Tóth F. <janos666@xxxxxxxxxx>:
> These are CCT and Y measurements on a NEC PA231W with the uniformity
> compensation set to 0 and 5 (off, max):
>
> img802.imageshack.us/img802/7595/pa231whomogeneitycomp.jpg
>
> But you obviously loose a lot from the contrast ratio of you use this feature.
>
> 2012/4/16, Graeme Gill <graeme@xxxxxxxxxxxxx>:
> > Alexey Blinov wrote:
> >> I measured color on different places on a fully white screen. When
> >> visualized on the CIE 1931
> >> x,y plane, I got an area of size around 0.01. The distance between two
> >> closest shades of white
> >> (255,255,255 and 254,255,255) is around 8e-4. Accuracy of the colorimeter,
> >> i.e. the diameter of
> >> the smallest circle enclosing many measurements of same color at same
> >> place, is around 4e-4.
> >>
> >> So, variations of color on the screen is around 10 times more than the
> >> variation due to limited
> >> (8-bit per channel) digital representation. Who measured such unevenness
> >> of professional
> >> monitors, could you post your results?
> >
> > Hi,
> > spatial color unevenness in screens is a quite well known display
> > defect.
> > One of
> > the specifications of higher quality screens is improved spatial screen
> > uniformity,
> > and many of these displays take technical measures to ensure uniformity,
> > such
> > as spatial calibration systems.
> >
> > (I'm not sure x,y values are a good way of evaluating non-uniformity, since
> > they
> > do not correlate to visual differences. Delta E is a better measure in this
> > regard.)
> >
> > There are various industry standard methods and thresholds for display
> > uniformity
> > measurement, such as VESA and ISO standards.
> >
> > Graeme Gill.
> >
> >
Other related posts:
- » [argyllcms] Re: [argyllcms] Re: Monitor screen color unevenness/irregularity - Alexey Blinov
