Digital BW, The Print

> From: Bill Morse [mailto:willym@...]
>
> When you scan large numbers of negatives, and in particular older
> negatives,
> you learn to expect subtle differences in the channels, from many
> different
> sources, and to make adjustments accordingly.

That may be true, but I would expect that on a particular negative, the
tonal curve would be the same everywhere, so could be corrected to whatever
you want with a single global Curve. But I defer to your experience. I've
just never seen anything in a B&W neg that looked like even faint color.
It's been years, though.

> I have heard, but cannot personally confirm, that PS as well as many
> scanning programs over-emphasize the red channel over the green and blue
> ones, supposedly to \ufffdbetter render\ufffd (read emphasize) Caucasian skin tones.
> I seem to remember something like 50 red, 30 green, and 20 blue
> (don\ufffdt quote me on these exactly! ;^)

I tested it using a gamma=1 color space, and Image->Mode->Grayscale weighs
them 80, 160 and 15 (out of 255), so it's green that gets the strongest
emphasis. However, the default settings in Channel Mixer are 100,0,0, which
I suspect is to make blue skies dark.

> In my experience, most RGB scans of a B&W source will show
> greater noise and
> more emphasized grain in the red and blue channels, so I typically start
> with an emphasis on the green channel, say 30 red, 45 green, and
> 25 blue.  I
> find that as a starting point, this will give me less noise and minimal
> grain, as opposed even to using 100% green.

Yes, I'd expect green to be quieter, because it captures more light energy.
I wonder why you get more noise if you go to 0,100,0. Or is some of the
noise from film grain, as opposed to CCD noise?

--

Ciao,               Paul D. DeRocco
Paul                mailto:pderocco@...