Re: Media Testing for maclife.de
Re: Media Testing for maclife.de
- Subject: Re: Media Testing for maclife.de
- From: Eric Chan <email@hidden>
- Date: Thu, 11 Sep 2008 05:53:16 -0700 (PDT)
Yes, the CFA design largely accounts for how well camera response can align with human vision or (possibly more interesting for some users) how well two different camera models can produce the same color. If CFA responses were simply linearly related to XYZ primaries, then a simple linear transformation (i.e., 3x3 matrix) would suffice to perform a "perfect" colorimetric characterization of the sensor.
Unfortunately, I am not aware of the existence of any such CFA. Some cameras exhibit very high IR sensitivity which is a nightmare for skin tones. (Skin has high IR reflectance, which is why IR-modified cameras render skin tones very light; same with foliage.) Shooting any standard color target, no matter how many patches, largely fails to account for this. An optimized color matrix (regardless of which optimization metric you use) is likely to lead to excessively pink skin tones, particularly under tungsten illumination (high IR energy). This is just a case of the camera "seeing" differently than we do, and profile tweaking is needed to make the skin tones look better (even though the nasty-looking pink rendering is arguably more colorimetrically accurate from the camera's point of view).
This is less of an issue with some of the more recent DSLRs.
My point about the inkjet printing was simply that the colors produced are dependent on the spectra of the (very limited number of) pigments, or perhaps I should say, pigment+paper combo. So you could, in principle, print a gazillion patches, but after you get beyond a dozen, you're not actually getting any new information.
Eric
--- On Thu, 9/11/08, edmund ronald <email@hidden> wrote:
> My feeling is that what realy makes or breaks camera color
> is the
> response of the Bayer CFA in the camera - which is clearly
> out of the
> hands of Adobe or anyone else. If the CFA is badly matched
> to human
> vision or has spikes then a calibration may work on a
> target but that
> camera will make its users unhappy. Has happened.
> If the CFA matched (is that possible ?) then calibration
> with almost
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden