Going through serious gum withdrawals so I spent a couple of days reliving the agony. I’ve been going round and round with the separation technique, but finally proved to myself the multichannel and RGB give identical results (at least on CS4), while the CMYk ( K layer as needed) method tends to give more saturated isolated colors. This duck image illustrates that pretty well in the yellows, oranges and reds- note the duck body and bill. The channels are in order left to right- C, M, Y. I do print the K channel with ivory black, to build the shadows back up.
Should note a few things-it’s especially important to convert rgb to cmyk before inverting the image to a negative. That will keep the K channel info in the shadow of the print instead of muddying up the highlights. Also it will probably be necessary to play with the white points of the negative a little (the shadows of the print), depending on how much you want to rely on a final K pass.
For clarity, the bottom row is the original file, the rgb version print, then the cmyk version print. Clearly both prints have issues (they’re are among my first attempts) but if nothing else they do illustrate the worst of each approach. I’m aiming for somewhere between the RGB and CMYK. The CMY(and K as needed) method sacrifices dimensionality and depth in favor of saturation, and the RGB sacrifices exactly the reverse.
Update 9/28
Following a recommendation on the hybridphoto forum (thank you Loris), I started messing around with the saturation levels pre-separation and just using RGB. Here are the extremes so far- +75 saturation on top, plain RGB on the bottom. Next, I think I will selectively saturate the colors that I need- just really wanted to bump the dry green and cyan of the juniper tree here, and leave the canyon wall subdued. Should be easy to do by bumping the yellow, green and blue a little, and maybe desaturating the reds and magentas a little.
Update 10/9
I’ve been neglecting screen proofing, which might be a grievous error. Turns out setting up the custom proof for working CMYK with simulate paper color on actually gives a pretty good proof of the desaturated RGB results I’ve been struggling with. So it’s much easier to tweak the saturation and contrast up with the screen proof on, and get the desired effect in the final print, rather than just guess at an arbitrary saturation boost.