Creating a CCMX profile

[Yuri D'Elia]

So I've got my hands on a Gretag Macbeth i1 monitor (emissive only).
Could I use it to create a CCMX profile for the colorhug?
If so, is there a guide I can follow?

[Yuri D'Elia]

I'm dumb, since I've eventually found this page:

http://www.hughski.com/calibrate.html

colorhug-ccmx doesn't recognize the i1 monitor directly.

I've actually used dispcalGUI to make the process a little less tedious
(tools -> create colorimeter correction). I've set the colorhug to 'raw'
mode when reading the testchart (which uses dispread -yR). Anything
wrong with that as opposed to resetting the hug?

What about the display backlight? Should the panel be set to maximum
brightness when creating the ccmx? I've set mine to roughly 130cd (45%
brightness).

This is what I get when using the i1+colorhug against an HP 840 G1 laptop:

Fit error is max 0.389602, avg 0.303353 DE94
Correction matrix is:
1.356377 0.654106 0.479147
0.595945 2.433718 -0.452696
-0.456809 -1.177274 4.803367

Is there anything I could do to test the created ccmx?

[Yuri D'Elia]

On 04/08/2015 08:29 PM, Yuri D'Elia wrote:
> Fit error is max 0.389602, avg 0.303353 DE94
> Correction matrix is:
> 1.356377 0.654106 0.479147
> 0.595945 2.433718 -0.452696
> -0.456809 -1.177274 4.803367
>
> Is there anything I could do to test the created ccmx?

I should have added "in an objective way" since, I guess, I could just
compare the two profiles (i1 / colorhug) against each other.

I'm working in a large institute and I could create correction profiles
for several displays here. Before uploading though, I'd like to make
sure the ccmx was correctly generated.

[Richard Hughes]

On 8 April 2015 at 19:29, Yuri D'Elia <wav...@thregr.org> wrote:
> colorhug-ccmx doesn't recognize the i1 monitor directly.

Ohh? Is that with a newish version of colorhug-client?

> I've actually used dispcalGUI to make the process a little less tedious
> (tools -> create colorimeter correction). I've set the colorhug to 'raw'
> mode when reading the testchart (which uses dispread -yR). Anything
> wrong with that as opposed to resetting the hug?

That should be the same exact thing.

> What about the display backlight? Should the panel be set to maximum
> brightness when creating the ccmx? I've set mine to roughly 130cd (45%
> brightness).

Yes and no. Max brightness might subtly change the color, but it's
safer to do that than risk the ECO mode causing flicker which confuses
the poor L2F sensor.

> Fit error is max 0.389602, avg 0.303353 DE94
> Correction matrix is:
> 1.356377 0.654106 0.479147
> 0.595945 2.433718 -0.452696
> -0.456809 -1.177274 4.803367

That looks pretty large; is this a wide gamut display?

> Is there anything I could do to test the created ccmx?

I normally do a quick sanity check using gcm-picker.

Richard

[Yuri D'Elia]

On 04/12/2015 08:26 PM, Richard Hughes wrote:
> On 8 April 2015 at 19:29, Yuri D'Elia <wav...@thregr.org> wrote:
>> colorhug-ccmx doesn't recognize the i1 monitor directly.
>
> Ohh? Is that with a newish version of colorhug-client?

With 0.2.1 as before, I'll try again with an updated version.

>> I've actually used dispcalGUI to make the process a little less tedious
>> (tools -> create colorimeter correction). I've set the colorhug to 'raw'
>> mode when reading the testchart (which uses dispread -yR). Anything
>> wrong with that as opposed to resetting the hug?
>
> That should be the same exact thing.

Excellent.

>> What about the display backlight? Should the panel be set to maximum
>> brightness when creating the ccmx? I've set mine to roughly 130cd (45%
>> brightness).
>
> Yes and no. Max brightness might subtly change the color, but it's
> safer to do that than risk the ECO mode causing flicker which confuses
> the poor L2F sensor.

I see.

>> Fit error is max 0.389602, avg 0.303353 DE94
>> Correction matrix is:
>> 1.356377 0.654106 0.479147
>> 0.595945 2.433718 -0.452696
>> -0.456809 -1.177274 4.803367
>
> That looks pretty large; is this a wide gamut display?

It is (dell u2413).

What are the expected ranges for a normal lcd monitor, so that I can
quickly validate the results?

>> Is there anything I could do to test the created ccmx?
>
> I normally do a quick sanity check using gcm-picker.

Could you elaborate a bit please?
Since I never did this before, I would need some guidance.

[Richard Hughes]

On 12 April 2015 at 21:23, Yuri D'Elia <wav...@thregr.org> wrote:
>> That looks pretty large; is this a wide gamut display?

> What are the expected ranges for a normal lcd monitor, so that I can
> quickly validate the results?

Actually, maybe I was a bit hasty, that could actually be correct if
you scale things to unity. Once scaled, anything much over 1.0 makes
me raise my eyebrows.

> Since I never did this before, I would need some guidance.

Run gcm-picker, put the ColorHug on a #ff0000 and see how well the
colors match; they're never going to be 100% accurate (and really
doesn't have to), but it should be a lot more accurate than with the
factory matrix.

Richard.