[argyllcms] Re: Displaying sRGB graphics on wide gamut monitors - gamma problem?

  • From: Anders Torger <torger@xxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 27 Oct 2010 17:50:02 +0200

Thanks for that thorough information, really helpful!

It actually seems like I got it kind of right altough my terminology is 
not so precise. I did think that calibration is getting the screen "back 
to neutral" so it behaves nice and well within its color space (and 
adjusting whitepoint, brightness, gamma, that was what I meant by 
desired target), and profiling is measuring and modeling what that 
neutral really is.

However, I need to study more how the data in these tables are actually 
used. The source of all this was that I expected vcgt values at 0,0,0 to 
be 0,0,0 with the -k 0.0 parameter to dispcal, but clearly that is not 
the case and now I need to find out why (or go mad, or both).

/Anders


On Wednesday 27 October 2010, Sam Berry wrote:
> On 27 October 2010 15:21, Anders Torger <torger@xxxxxxxxxxx> wrote:
> > On Wednesday 27 October 2010, Graeme Gill wrote:
> > > Please take some time to understand the difference between
> > > calibration and profiling.
> > 
> > This is what people keep saying to beginners like me :-). However,
> > it seems to me quite easy. Calibration = through translation
> > (matrix or lut) make the monitor display colours closer to some
> > desired target.
> 
> This is profiling.
> 
> > Profiling = measuring the screen to see how it actually does
> > display colours. One can profile a calibrated display or an
> > uncalibrated display.
> 
> This is profile/calibration validation.
> 
> Calibration = Ensuring that display channels perform *individually*
> as expected. Normally via display settings and 3 1-dimensional LUTs.
> This does not alter colour in a major way, other than perhaps
> altering the white point. A perfect monitor will show no change
> after calibration. This process ensures the monitor conforms to a
> 2.2 gamma response with a D65 white point, for example.
> 
> Profiling = Modelling the display using a matrix or a set of 1D LUTs
> and a *3-dimensional* LUT. This is the step that allows the graphics
> application to ensure the display shows the correct colours,
> regardless of varying gamut. Prior calibration improves the
> effectiveness of profiling.
> 
> Validation = Ensuring that the final result of both of these steps is
> adequate by measuring a set of patches and comparing to those
> predicted by the profile.
> 
> > However, a source of confusion may be that I have misunderstood
> > where the data is. I thought that for an .icc profile from dispcal
> > vgct = calibration curves and rTRC/gTRC/bTRC = profile (*after*
> > vcgt applied)... but I'm no longer that sure. I must try to look
> > further into this, source code if I must. It will be impossible
> > for me to find the source to my experienced (or imagined) problems
> > if I don't have full understanding how calibration and profiling
> > is applied.
> 
> The xTRC curves are there to allow the colour correction to be
> performed in linear space, rather than in gamma space. Matrix gamma
> transformation operations are meaningless unless performed on linear
> data. The curves allow this transformation and its inverse to be
> performed.
> 
> > Anyway, I'm sorry for being clueless, I know it's not fun educating
> > clueless beginners like me... I'm trying not to say clueless
> > things, but it seems today like I need to spend some more time on
> > this before leaving the clueless stage :-).
> > 
> > /Anders
> 
> Sam Berry
> www.satsumatree.co.uk


Other related posts: