Manuel Viet wrote:
>Le vendredi 03 Mars 2006 01:51, Moose a écrit :
>
>>I do my image editing in PS,
>>CS/8 for the moment, but it doesn't matter which version for this. I've
>>made a little step by step example of what I did to your image
>><http://www.moosemystic.net/Gallery/tech/Tut01bw.htm>.
>>
>>
>I understand you begin by scanning without telling us anymore details
>
Not me. This is Jonas' image. He posted it, I posted positive comments
on subject & composition and some critique of tonality, etc., along with
some examples of how I might want it to look
<http://moosemystic.net/Gallery/Others/Sum05.htm>. He posted a link to
the original scan file and asked for some info on what I did to it, so
I posted the little step-by-step. You have to ask him about such
scanning details.
>to add my own grain of salt, I'd advise to scan b&w pictures in color mode,
>then use the in-software b&w converter. Rationale being that I have a scanner
>which scan only in the red channel while in b&w mode; maybe yours don't
>
I don't know whether mine does or doesn't. It shouldn't matter. For one
that does, and I suspect most do, your solution sounds right only if
there is something wrong with the scanner. A true B&W source should
produce the same results no matter which channel is used.
>, but it can be frustrating to see a dark image from an otherwise correct
>print.
>
This was a scan from film, not from a print.
>Scanning in color then turning to b&w made this problem vanish (to the
>expense of your RAM - better have a lot of it anyway).
>
>The added benefit is split-adjustment of every rgb layer before collapsing the
>picture into b&w. This has the same effect as putting a colored filter on the
>lens, but you can choose your exact filter as an afterthought "in lab", very
>convenient.
>
>
Certainly possible when scanning from a color source, although tere are
other methods I prefer. When scanning from a B&W source, a properly
operating scanner should produce identical data in each channel, so
adjusting individual channel brightness should have no relative effect
on tonality between subject colors. It may change overall brightness,
but not those things affected by color filters in taking a B&W image.
>The more I use scanners, the neater I find them ; they have an incredible
>"view", sometimes getting details out of a photopaper print you couldn't
>possibly imagine to have - invisible to the naked eye (and I have a very very
>good eyesight, far above what's considered "normal good").
>
Yes - and no, relative to what's on the film. And it depends of course
on the size and quality of the print. A reflective medium simply can't
display the dynamic range of a transmissive medium like film. So a scan
of a print must always do some combination of compression of the range
and loss of highlight/shadow detail. Done properly, it can be very good,
but nonetheless cannot reproduce all the tonal range/detail on the film.
The great B&W historical landscape/art photographers often used dodging
and burning extensively to change the tonal relationship of various
parts of the image at least in part to deal with this limitation of prints.
Also, there is always some loss of detail. No matter what enlarging lens
is used and no matter how big the print, the lens will be less than
perfect. The amount of loss can be very small with great care and large
prints. With smaller prints and less perfect lens and technique, it can
be considerable. Of course, the same is true of film scanners, but the
best seem to capture detail beyond most prints.
Some time ago, a list member kept complaining about the crummy scans he
was getting from his color prints done on great commercial equipment,
and that looked great to his eye. I posted this comparison of flatbed
scan of a Kodak Master, or some such name, print from their quality
processing division, not the drugstore stuff, and a 2720 dpi film scan.
The comparison of dynamic range, detail and sharpness is quite an
incentive to avoid such prints as a source
<http://www.moosemystic.net/Gallery/tech/PrintvsScan/ScanvsPrint.htm>.
I've since scanned it at 4000 dpi on a newer scanner; not a great deal
of improvement. A little less obvious grain and a tiny bit more detail,
but nothing that would show up in anything but an enormous print.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|