[argyllcms] Re: Monitor calibration
- From: Steffen <mail@xxxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 24 Mar 2010 09:55:53 +0100
Am 24.03.2010 08:43, schrieb Pascal de Bruijn:
> On Wed, Mar 24, 2010 at 1:04 AM, Steffen <mail@xxxxxxxxxxxxxx> wrote:
>> There are indeed monitors with unequal RGB gain values from factory,
>> like the Eizo S2433. Whatever works best, I suppose.
>>
>> And there is a distinct advantage when working with the gain controls on
>> the monitor to achieve a starting point as close to the target as
>> possible. Most of the higher grade monitors have a 10 bit deep LUT. That
>> means, that unlike the 8 bit available through ordinary LUTs for the
>> software correction on the graphics card, changing values in the display
>> won't "throw away" as much information as a correction on the software
>> side would. I hope this is what the discussion was about, if not, I
>> apologize for my lazy reading habits ;-).
>
> Higher grade being 1000EUR+ :) ??
>
> Anyway correcting in two places in theory degrades more, no matter how
> many bit the LUTs are...
>
> Regards,
> Pascal de Bruijn
No, most PVA or IPS displays feature a LUT that is 10 bit deep. Only
most TN panels don't. So you don't have to sacrifice all your money in
order to get some bang for the buck. I wouldn't be surprised if the HP
LP2475w and the Dell U2410 also had such features.
And in general terms, I would agree with your assessment. But not
necessarily this time: if you calibrate based solely on software
correction, especially if the deviation from the target is big, your
results WILL definatelly be worse. Because if you correct in hardware
first, 10 bits result in 1024 steps, compared to the 256 available steps
at 8 bit. So, you have more room to operate in without touching the last
8 bit your graphics card will send to your display.
I think this is definatelly preferable.
Other related posts:
