Digital BW, The Print

This is very interesting, indeed. It makes sense, though, that soft 
proofing works. A profile contains two mappings, one from the color 
connection space (PCS) into the printer space (if you want a 
particular color, what set of RGB values should be sent to the 
printer), and one from the printer space to the PCS (for each RGB 
value sent to the printer, what color is produced). 

When you generate a profile using x-tone inks, the profile should 
accurately represent the 2nd mapping, from the printer space to the 
PCS. In other words, for a given RGB value sent to the printer, the 
profile will know what grey tone is produced, since it measured it on 
the target.

When you soft proof, Photoshop figures out each RGB values it should 
send to the printer for the set profile and rendering intent, then 
uses the profile to figure out what color the printer should actually 
produce. 

The tricky part with an x-tone profile is that the gamut of the 
printer space (tones of gray) is much, much smaller that that of the 
working space (such as Adobe RGB), so the success of a profile really 
depends on how the profiler handles out-of-gamut colors. Apparently 
Profile Prism handles them gracefully. A bad profile would freak out 
when, say, bright green needs to be printed. a good profile would try 
to squeeze it sensibly into the x-tone color space. For a Roark 
workflow, the best situation would probably be to pass the RGB values 
in the image directly to the printer, without color management, and to 
use preview to see what the printer will produce. Passing the RGB 
values would preserve the output part of the Roark workflow. That 
seems to be your workflow as well.

A better way to approach this would be to produce a custom profile 
that does a partitioned PCS to x-tone mapping. For example, the 
lightness (grey value of the image) could be carefully mapped with 
proper ink partitioning, but the amount of tinting could be controled 
by the a- and b- channels in the L*a*b* space (all that means is that 
the color of the image would render the image cooler or warmer, but 
the luminosity would determine how the gray inks are partitioned).

Sorry if this was a bit technical and long.

FYI, I have made just such a set of ICC profiles for the MIS-FS 
Neutral inks on an 1160 printer on Enanced Matte and Crane Museo 
papers. I had to write my own software to do it, but it finally works 
well. I use a fully color-managed workflow and get very good, 
consistent results, with highly accurate previews. I can even mimic 
color filers, and see the effect on-screen! The inks are partitioned 
to minimize the dot patterns. I imagine the profiles are similar to 
the profiles produced by the Piezotone guys, but I found out about 
their efforts after I was nearly done, and they don't seem to have 
anything for MIS inks. 

I posted a very rough version of the profiles earlier in the year. If 
anyone wants the profiles and instructions I can get them posted in 
the Files area. Alas, I doubt that many people are using MIS FS 
Neutral inks on an 1160 anymore, with these fancy Ultratones...

I think I can profile just about any ink set, including Ultratones, 
but it'd be months before I could get to it because I'd have to modify 
the software to do full-color previews (right now it just does 
previews of lightness, which works fine for a neutral inkset). If 
anyone wants to contact me off-group, we can discuss it.

A patchwork approach might be to try to imbed Paul Roark curves into a 
profile, while using the profiler to generate the preview. That I 
might be able to do faster. 

- Steven


--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Editor P.O.V. 
Image Service" <editor@p...> wrote:
> Ok, some years back when the VM-S set hit the streets, I wondered 
aloud 
> on this list about profiling it with a commercial profile package..
> 
> It created a stir, and I tried doing it with some then commercially 
> available packages... Results were VERY disappointing and resulted 
in 
> huge amounts of posterization.
> 
> Since then I've played with a bunch of products and had been having 
> surprising success on color profiles with ProfilePrism of all 
things..  
> I never even thought to try a "low-end product" like Profile Prism 
on 
> the previously uncompleted task of properly profiling the VM-S set..
> 
> Well, I've been a tad bored lately and was running a bunch of other 
> tests, creating actions, etc.,  So, I figured, "what the hey, why 
not 
> give Profile Prism a shot at profiling the VM-S inkset?"
> 
> Amazingly, it created a profile that is both usable for printer 
> profiling and  soft proofing (a necessary component to getting this 
> right)..  The workflow is pretty straightforward, but it works.. 
Color 
> me astonished. (I'm guessing that Profile Prism does more simple 
point 
> to point mapping and less interpolation as far as: rendering intent, 
> smoothing, etc. - meaning a profile for something like this would be 
> more likely to be usable)
> 
> While some of the output still needs a tad bit of tweaking in the 
> shadows (and that can be accomplished with a curve if necessary), 
this 
> workflow should allow split toning at will with the VM-S and VM 
inksets, 
> as well as perhaps the UltraTones...
> 
> If there's interest, I'll put together a complete 
workflow/explanation 
> and send it to Martin Wesley to add to the group files..
> 
> 
>  
> Keith Krebs
> 
> "Just some guy," caretaker of the Multiverse's largest EPSON printer 
> User Community (highly recommended by Vogon Poets and MegaDodo 
> Publications), at:
> http://groups.yahoo.com/group/EPSON_Printers/
> and  the Multiverse's largest Canon printer User  Community at:
> http://groups.yahoo.com/group/Canon-printers
> "For the rest of you out there, the secret is to bang the rocks 
together 
> guys"