Digital BW, The Print

on 10/28/01 1:37 AM, Cleavis wrote:

> BACK TO THE QUESTION given this 'info'...When I bring something into
> photoshop 5.2 IT won't let me utilize most of the controls other than
> levels in 16 bit mode...so if you must immediately convert to 8-bit
> to manipulate why bother scanning higher??? Especially if chances are
> 256 greys are going to do it any way since the printer range is maxed
> out?

Cleavis,

This is a big topic, one open to much debate, so don't expect definitive
answers. 

The deal is this, and I'm gonna be brief (Austin jump in when I get it wrong
;-). Scanners really only scan one way, and that is raw. They read your
analog data and convert it/write it out as linear digital code. This results
in a raw file, which when you see it as an image, it's dark and tonally
compressed (flat). To adjust it from this raw state requires software
manipulation of the data/image. This happens after the raw data has been
written. Your scanner's software (through it's driver settings) can do this
for you somewhat behind the scenes, and when your image opens from the
scanner it will already be adjusted. Or you can bring the raw data into
Photoshop and do the adjustments yourself - which is how I typically work.
The reason I do it this way is control. With PS you can view your data/image
at any magnification you want, crop, rotate, etc. and with history and
whatnot, undo mistakes.

Okay, to your point. There are a good many more controls available to you in
PS in 16-bit mode than just Levels. You can do Curves and Selections and
several filters. The big thing you can not to is Layers, and that IS a big
shortcoming, but there are workarounds. A recent issue of PEI magazine had
an article by Bruce Fraser where he shows how you can dupe the file, convert
the dupe to 8 bit, work your masks and layers on the 8 bit file, then load
those adjustments onto your 16 bit file. However, this is obviously more
time consuming, but it's very doable.

I split the difference. My method is safe and convenient (I'm used to it
anyway), but it's drawback is the need for huge hard drives or other storage
capacity. 

I capture my raw highbit data and archive that untouched. But first I dupe
it. The dupe I work globally (without selections) using Levels then Curves
in 16-bit to about the same point that I'd get by using the scanners driver
software. So now I have a nicely adjusted file that is tonally adjusted at
least as well as what the scanner would give me in 8-bit, but I have it in
16-bit. I dupe that then save it. The dupe of that I now convert to 8-bits
and start working that with selections and curves and sharpening; whatever I
want, anything and everything, and save that with all layers intact.

So now what I have is three files for this image: the raw scan, the 16-bit
globally adjusted file, and the 8-bit finished file. I keep these three
together in a folder of their own, and they are each somehow sensibly named
to reflect their mode. That there are three versions per image is why I need
lots of storage.

With this, I scan once, and can always go back and start over from scratch
with the raw file, or either adjusted file. The thing is, if I ever get to
where I feel that the local adjustments I made on my 8-bit file were
destructive to the image, and would have been better done in 16-bits, I can
load those adjustments, layer by layer, from my 8 bit file onto my adjusted
16-bit file, and with that I have the ease and control available from 8-bit,
with all the tones available from 16-bit. So it's very safe and flexible.

The truth is I've yet to find it worthwhile to load my 8-bit adjustments
onto my 16-bit image, I just stay with the 8-bit version. But I reserve the
ability to do it if changes to image processing (like more radical ink
separation curves are utilized) or better print output, make it more useful
in the future.

Confused yet?

Todd