Archived posting to the Leica Users Group, 2000/03/21
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]
2000-03-19-12:09:16 Mark Rabiner:
> A VERY interesting post ruined by this "wishful thinking" last
> paragraph, Mike. Showing a stack of prints to people is better
> "research" than the study of MTF graphs?
Well, yes, if aim is to determine which of two methods makes higher
quality pictures, and the definition of `higher quality pictures' is,
`pictures which look better to viewers' -- which I'd argue that it is,
in any context meaningful to this discussion.[1]
Improving how the pictures look to human viewers is the absolute,
bottom-line end goal. MTF curves are an extremely useful way of
representing aspects of the behavior of an optical system, and are
presumably an invaluable tool both for optical designers who wish to
be able to do repeatable rigorous comparisons of designs, and even for
potential purchasers who wish to try to predict results they might get
from equipment they can't try out. But they're a tool which is
valuable only to the extent that it aids in the production of pictures
which look good to people. They're a means to an end, not an end in
themselves.
One sees exactly the same conflict come up when people get too nutty
about trying to measure audio equipment. As long as they maintain
some perspective, and recognize that measurements of equipment meant
to reproduce sound for the entertainment of humans are useful *only to
the extent that the measurements correlate with what humans observe*,
all is well. But some folks insist on believing that the measurements
can be some absolute measure of quality, whereupon they're lost.
However precisely measurements are made, an observer of the last few
decades' developments in audio design will have noted that there's been
continuing progress in figuring out *what to measure*. The insecure
consumer who bases a purchase on, say, a THD number in a brochure rather
than on a personal audition of the equipment considered deserves the
earache he's likely to get.
[1] So we're not talking about reconnaissance photos, for example,
which *might* have different quality criteria.