> The problem, Cory, is that the question I was responding to, and
> subsequent discussion, was a bona fide question by a new user on how
> best to work with his K10D RAW image files. Not all of us are
> interested in theoretical pedantry.
>
My bad... I don't recall the original inquiry. I was originally
responding to technical issues of quantization and the "least amount of
data loss" in the image processing chain.
> There is no "truth in advertising" when it comes to taking a digital
> capture and transforming it to a quality rendering. Mathematical
> algorithms functions do exactly what they are programmed to do, to
> the limits of resolution of the representational medium. One works
> with the tools available to produce a pleasing/accurate (take your
> pick in whatever proportion you desire) result, regardless of what
> "errors" (more precisely, "deviations from theory") might exist.
>
> Pedantry does not produce photographs.
>
True. I am quite in the minority as a "technical photographer" as
opposed to an "artsy photographer." Most are the latter and whatever
looks good is acceptable. I find it difficult to trust my own sense of
quality, so I resort to objective means that I can quantify to produce the
most colorimetrically accurate and least manipulated images possible.
If I choose to apply a filter (blackpoint, whitepoint, curves,
levels, usm, greyscale conversion, etc, etc) to the image, I want the
operation to be conducted as accurately to the namesake as possible. It's
when they are incorrectly labeled that it bothers me. For something as
major as linear vs. log, its a huge deal from a data processing
standpoint... even if the end result can barely be detected.
I am cursed by the fact that I do not trust my own preference for
"pleasing/accurate," so I err on the side of accurate. I also understand
that most feel the opposite and are appalled at the pedantry.
To each their own.
-Cory
--
*************************************************************************
* Cory Papenfuss, Ph.D., PPSEL-IA *
* Electrical Engineering *
* Virginia Polytechnic Institute and State University *
*************************************************************************
--
PDML Pentax-Discuss Mail List
PDML@pdml.net
http://pdml.net/mailman/listinfo/pdml_pdml.net
