Thread-topic: Monitor profile verification [was: Eye One Pro for monitor calibration? [was: Re: NEC 2690 SpectraView]]
User-agent: Microsoft-Entourage/11.3.3.061214
In a message dated 12/2/07 7:14 AM, Roger Breton wrote:
> Terry,
>
> You'll agree that there is more than one way to skin a cat. Personnally, the
> many ways of evaluating a monitor's profiled bahavior are presenting facets
> of a same reality. Starting with RGB numbers is OK since they'll eventually
> be converted to Lab of some sort. It gets quite complex, actually, when you
> start looking at the conversions the data goes through on its way to the
> final numbers. I agree the gist of this evaluation largely depends on what
> one wants to evaluate. Starting with Lab is quite valid too.
>
> Me, I'm pluralist ;-)
Hi Roger.
It would seem that starting from an L*a*b* list in an 8-bit environment
could introduce a larger rounding error in the RGB values sent to the
display.
Sending even RGB numbers (R 255, G 255, B 128) in the "RGB-first" method, on
the other hand, seems to me to lessen the compounded amount of rounding
errors in the procedure. This compared to sending fractional values (e.g., R
234.78, G 135.23, B 67.98) in the "L*a*b*-first" method.
But perhaps none of this matters much, in the end, since the error is
probably very minute.
Marco Ugolini
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/colorsync-users/email@hidden
This email sent to email@hidden
