[Openicc] Changing the display whitepoint using the VCGT

[Chris Murphy]

I will walk back a bit on the "only Apple would know how to do" with respect to setting the vcgt. True, they have particular information that they know about this display, which is how they automatically generate an ICC profile with such a radical (and unique for each model of laptop, possibly even for each display type within a particular laptop as this Macbook Pro has three panel options). However their visual calibration app doesn't just let you move sliders so the software knows the native response of the display, it also has a hue neutralization feature for three or four points along the tone response curve. The idea is the user tries to make these areas of the TRC neutral, visually. So that's a way to end up with a non-linear vcgt to try and correct for color temperature not just at the white point, but throughout the TRC. And I think that's important because at least on this particular display, the white point is fine even with a linear vcgt, but tones below that, especially 1/4 tones are distinctly blue.

Chris


On Apr 26, 2011, at 4:27 PM, Chris Murphy wrote:

> I'm guessing you mean more like 6000K+ whites if it's bluish, or 7000K whites if very bluish. 5000K is "yellowish" to most people.
> 
> I would categorize vcgt adjustment of the white point "unfortunate" but also a reality if it needs to be changed, that's your only option on these displays. Ideally the panel manufacturer would use analog adjustments to ensure a reasonable white point for the panel+backlight source used, so that video card LUT adjustment needed it minimal. So I'd say maybe not do it by default because humans do have a chromatic adaptation feature, but keep it as an option for those who expressly ask for white point adjustment? If you pull back on the blue channel, for example, you get a better white point, but then also get a luminance drop. And that luminance drop can be significant and a worse effect than a non-ideal white point.
> 
> But I think it's going to depend on the display. I look at my new Macbook Pro 2011 (high res matte screen) with a non-linear vcgt produced by the OS, which is something only Apple would know how to do, and with another profile that has a linear vcgt. With the linear vcgt the display is distinctly blue. I don't know that I'd adjust to this, maybe after a day or a week? Certainly not in a few minutes trying it out. The white point is OK, but tones below the white point (grays) are definitely very blue.
> 
> I don't know why Apple does it this way, rather than analog adjustments to the panel to make the entire TRC relatively neutral.
> 
> Chris
> 
> 
> On Apr 26, 2011, at 8:48 AM, Richard Hughes wrote:
> 
>> Dear list,
>> 
>> My Dell mini display is very bluish by default (5000K whites), but
>> when I calibrate this using argyll and a hardware calibration device
>> the new profile sets a white temperature of 5700K and the display
>> looks a lot more pleasant. A very blue native white is also a problem
>> in some macbooks, and nearly all thinkpads.
>> 
>> I can easily create an RGB generated space using the EDID chromacity
>> co-ordinates and the EDID gamma, and gnome-color-manager has been
>> doing that for a few releases now. I'm wondering if it's sane to also
>> adjust the whitepoint using the VCGT feature.
>> 
>> I've been playing about with cmsAdaptToIlluminant and using the VCGT
>> which obviously does a linear conversion of white temperature, which
>> seems to work fairly well. I've noticed that argyll-generated profiles
>> seem to smooth to the maximum value a the top of the VCGT table and I
>> wondered if that was a result of manual tweaking to ensure that
>> 255,255,255 == display-white or may just be a peculiarity of my
>> display.
>> 
>> I guess the real question is whether trying-to-be-helpful session
>> software should try to change the whitepoint of a display to anything
>> other than native or whether the VCGT table should just be left
>> unmolested.
>> 
>> Advice very welcome,
>> 
>> Thanks.
>> 
>> Richard
>> _______________________________________________
>> openicc mailing list
>> openicc at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/openicc
> 
> _______________________________________________
> openicc mailing list
> openicc at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/openicc