Thanks to Dr. Exposition. If he hits me over the head a few more times
with "look at the histogram" it may finally sink in.
Dr. Flash
Moose wrote:
> Chris Barker wrote:
>> I'm sure that that's a beautiful shot, Graham, but colour looks wrong on my
>> monitor. It seems always to be the case with poppies when I take photos:
>> digital does not like that colour.
>>
> Chris, and others who have replied in agreement,
>
> This is a common problem with digital cameras and/or processing that is
> especially common with intense reds. The cause is clipping on one colour
> channel before others.
>
> You can see the effect in the individual channel histograms in a small
> section of the flower image here.
> <http://www.moosemystic.net/Gallery/tech/Misc/RedClip.htm>
>
> The original distribution of red in the subject was likely much like
> that of green, a narrow, symmetrical distribution. As colour neg film
> has a great deal of overexposure latitude, red layer brightness in the
> film probably pretty well matched the subject. Somewhere in scanning
> and/or processing, the red channel was overexposed, and the right side
> of the distribution all ended up with one value.
>
> One visual result of this is a lack of subtle detail in the red
> highlights. Another, as noted explicitly by Chuck and implicitly by
> others, is a change in colour. In the brightest pixels, the red can't be
> increased in proportion to the other channels, so color* casts appear.
> In this case, as the blue channel is brighter than green, the color cast
> is slightly blue. The reds in the central, shaded portions of the flower
> image are clipped much less, so the color is more likely accurate.
>
> The same thing can happen with film, but is less common and/or obvious
> because of the exposure characteristics of film vs. digital sensors. The
> softer shoulder of the film response curve doesn't sharply clip above a
> certain brightness level. In theory, I suppose it's more likely to be
> noticeable in slide film than color neg.
>
> I also suppose the same effect could occur on a monitor set too bright,
> so that a color channel clips where the others don't.
>
> Moose
>
> * Spelling of colo(u)r in this document is customized to the native
> spelling of the poster(s) to whom each individual instance refers.
>
> ==============================================
> List usage info: http://www.zuikoholic.com
> List nannies: olympusadmin@xxxxxxxxxx
> ==============================================
>
>
> ------------------------------------------------------------------------
>
>
> No virus found in this incoming message.
> Checked by AVG.
> Version: 8.0.100 / Virus Database: 270.4.1/1514 - Release Date: 6/23/2008
> 7:17 AM
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|