Digital BW, The Print

Mike,

here are some things to consider:

- If you are scanning in RGB anyway, it's a good idea to maintain a layered 
RGB file with all the corrections - including the channel mixer or whatever you 
use to make it look monochrome. By loading a gray profile in the Gray Setup 
that accurately matches your output, all you have to do is drop a copy of the 
file to grayscale and print. You rename that copy for the device/ink/paper etc 
it's meant for. Also you can do USM specific to the final print size at that stage.

The advantage here comes later: If any of your conditions change, you load 
up a revised gray setup and all you do is drop a new copy from the RGB 
layered to grayscale and voila, perfect match. You can then do a whole batch 
with near-automation (an action would do, too).

- The alternative, which is to convert to bw and then do a bunch of corrections 
on top (layered) means you have to be aware of what your gray setup is. If you 
are using a "generic" setup like gamma 2.2 which matches your RGB gamma 
(like Adobe 98), you are not twisting the data too much and you can do a 
profile-to-profile (and USM) to a copy of the layered file before you print in gs 
mode. That seems reasonable except for the effects of the calculations you 
perform on the image through curves, layer modes etc. RGB offers more 
options and perhaps a smoother histo at the end when you drop to bw (for 
reasons you already suspect). 

- Do a test for yourself. Create an RGB gray ramp (blend tool), layer it with 
corrections such as you are likely to use in your scans, than drop a copy to 
grayscale as I describe above. Look at the histo. 
Then do the same thing in grayscale, look at the histo. 

- As far as how the RIP sees the data, if all RGB values are the same among 
channels it makes no difference. If you plan on using "colorized" RGB data 
that's another story and would depend on the separations built into the RIP 
(but that wasn't what you asked).

It boils down to the specifics: Do you need heavy duty manipulations? Do you 
foresee printing the same image at a later time in a different print 
environment? Is it worth the extra space for the data to maintain a master file 
whose bottom layer is the actual scan? 

These count more than the  purely technical reason of the extra bits in RGB. In 
some cases there may not be a visible difference that's worth 3x the data.

What's your take on this?

Antonis

<don't need no stinking pocket protectors for my spectrocam! - (to Todd)>
<BG> 


--- In DigitalBlackandWhiteThePrint@y..., "mkravit" <michael.kravit@w...> 
wrote:
> Utilizing the Colorbyte Software ImagePrint 4 RIP I find that I have 
> the option of printing with guadtone (hex) inks as either RGB or 
> Grayscale image files.  That is, I can scan a negative in RGB mode as 
> most people do, convert it to b/w using the channel mixer, but 
> continue to save the image as a RGB file.