AG Schnozz wrote:
>I'm finding that some mild local contrast enhancement using USM
>is just what the doctor ordered. I agree with the "tooth" bit.
>Me thinks it has something to do with the straight-line section
>of the response curves.
>
I think so too. The curves built into DCs and scanners/software just
don't seem right for some subjects. Sometimes curve adjustment works
best, sometimes LCE, sometimes a bit of both.
I intentionally scan to preserve almost all of the brightness range,
which can result in some pretty flat looking images. With LCE, curves,
highlight/shadow, etc. I can then get the image qualities I want without
dumping the ends of the histogram before I even start. I may compress or
expand parts of it, but not chop them off. Working in 16 bit to do all
this before converting to 8 bit for display/print leaves a histogram
that is witout the holes and lumps you get doing it all in 8 bit.
>Playing around with the curves sometimes adds that bite that you're looking
>for.
>
Yup. sometimes a full S curve does it, sometimes just a tiny movement of
the center will just work wonders. All kinds of things can be done there.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|