Digital BW, The Print

Michael King <drmrking@...> wrote:
>
> ... because of ink position sharing I plan to get 
> "Tri Hextone" such as
> neutral / sepia / split out of 11 ink positions.
> 

That sounds interesting.  The general discussion brings up an interesting issue I've found in my experiments with profiling and the use of differing numbers of inks and types of profiles.

The dogma is that the more channels of gray ink we have printing, the smoother the final print.  This dogma is probably a good starting point, but my experience is that not only is it characterized by decreasing marginal utility, but more may actually be worse in some situations.

With the Epson 1800 and Eboni, in the experiments that ultimately lead to the "3-MK" workflow with 3 full strength Eboni channels, I tried all sorts of different numbers of channels of that ink.  For that printer and approach, the smoothest was about 3.5 channels.  That is, while subsequent experience found that a minority of the 1800 printers needed 4 to be smooth, mine was about equal between 3 and 4, and less smooth above that.  While the final decision to go with a 3-MK approach obviously was influenced by that number also allowing color to be installed, I found that more channels did not necessarily result in smoother prints, at least with that type of arrangement.

Similarly, with a good 1400 and using a channel that does not band (not all channels are equal on mine), a single black only can be smoother than  the 1800 3-MK workflow.  

At least with the use of multiple high density inks, as in the 3-MK approach, it appears that once enough are employed to hide the banding, the use of more channels begins to increase the noise in (roughness of) the print.  With scanning and digital exposures I'm accustomed to thinking that random noise is averaged out with multiple sampling.  With printing at least with the 1800 3-MK setup, however, noise appears to be additive.   

Note that this issue is where multiple channels are firing at the same time, not where they are used in a serial partitioning profile. But it does raise the issue of whether, using multiple overlapping channels will always lead to a smoother result.

While I've been a fan of making profiles with more overlaps in them, with my 7800 I tested profiles with overlapping, dual channels, using QTR, and found the smoothest was the single, serially-partitioned profile -- the standard, simple QTR all-gray (no "toners") profile, made using Roy's semi-automatic partitioning algorithm. 

It may be that there is a trade off between hiding banding and increasing noise in the system.  Also, I think there is an issue of the tolerance of a profile to inconsistent papers and inks, and that issue probably favors more channels firing at the same time.  

Overall, there seem to be a lot of different issues that affect our printing, but it may be that more inks firing at the same time is not necessarily a good thing.  Noise, including possibly not only the dots but the dithering patterns, in our printing systems may, at least in some respects, be additive.

Paul
www.PaulRoark.com