Digital BW, The Print

Bruce, I don't want to get into too much trouble here. Roy and others
know much more about the inner working of dithers etc..
Your assumptions below I believe to be correct. For 2880 the smallest
dots are used, for best settings with 1440 I believe all dot sizes are
used in every ink with the "intellegence" you mention. I'm not sure
what settings may insure only the largest dot in use, maybe 720?
So no, no bad.
I'm pretty sure this is how the Epson driver works. With SP, there are
many more controls as you know, but It's obvious only ultra micro is
available with 2880, many more options from single size to variable
when other quality settings and dithers are selected. For example,
7000/9000 output can be mimicked on the 7600/9600s by selecting
"normal" dot size, rather than another that will allow micro or more,
even though both are available for 1440x720 output.
You are also correct about unusual media. There are papers I play with
that clearly cannot take the largest dot size without mottle or bleed.
So some optimization is possible for those, like japanese.
It's possible to put down even more ink than just the largest dot
size, I'm not sure how it's done, but I can make HPR bleed like crazy
with some settings and achieve a 1.75 dmax. But it's unusable.
Don't know the answer to coverage vrs dot size and how it relates to
dmax, to both you and Greg.
I do know that the various generations of Epsons I've had, the larger
dot printers always had the highest dmax on good papers.
Tyler

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, hogarth@... wrote:
>
> Well, I didn't understand how the printer actually wants to work - that 
> is, given a printer resolution, a particular dot size is used by all 
> channels. Smallest dots for 2880, bigger dots for lower resolutions. I 
> thought Epson meant to use this feature for more than just that.
> 
> What I was envisioning (be careful there!) was a driver smart enough to 
> use all the dot sizes all the time, and smart enough to determine
when a 
> given dot size was appropriate. The drivers (Epson, StudioPrint, QTR, 
> etc.) seem to be smart enough to run dither patterns and decide how to 
> place dots, and which dots to place, to make good smooth tonal 
> transitions. I just thought they could do this with dot sizes too, 
> because clearly variable dot size has the potential to improve the 
> smoothness of tonal transitions, and improve sharpness too. Yet
variable 
> dot size doesn't appear to be used that way. My bad.
> 
> I can see how it would be good to have the biggest dots in the darkest 
> regions - to ensure good coverage of the substrate so that no "paper 
> white" shows through. As in, better Dmax. This would clearly be easier 
> to do with bigger dots since they could more easily overlap if
needed. I 
> can also see that the smallest dots would be most useful in the
lightest 
> tones because there you have to have some paper white showing
through to 
> get those 1% tones.
> 
> But I can also see how the ink and coating are going to have a major 
> influence on when and where you need to use the big or small dots. Dot 
> gain comes to mind. If you have a lot of dot gain, you probably don't 
> need the biggest dots to ensure full coverage of the darkest areas. If 
> you have less dot gain, you might need the bigger dots. That sort of
thing.
> 
> It's almost like we need to extend the ICC profiles - to carry 
> information like that to help the driver do a good job of picking dot 
> sizes as required to get the tonal transitions we want. Before that 
> though I want the ICC profiles extended to do a better job of handling 
> neutral tones ;-)
> --
> Bruce Watson
> 
> 
> 
> Tyler Boley wrote:
> 
> > --- In DigitalBlackandWhiteThePrint@yahoogroups.com, hogarth@ wrote:
> > >
> > > So... v12 doesn't have a working algorithm to pick which size dot to
> > use
> > > for each ink then? I would have thought that this is part of
what you
> > > pay the big bucks for. This information doesn't make me want to
lay out
> > > the money for the upgrade.
> >
> > Bruce, I'm not absolutely positive I understand what you mean by such
> > an algorithm but if I do, I think it would be extraordinarily complex.
> > It would be dependant on the nature of the particular ink, and the
> > particular media. There is obvioulsy a built in algorithm that
> > transitions from one dot size to the next, probably based on Epsons
> > own. You can control which dot size or sizes to use and how much of
> > each, per ink. But after playing with this a bit, I can't imagine how
> > it could be automated. Too many things have to be seen to decide how
> > to proceed, and for mono inks, I'm seeing no need for the larger dot
> > sizes at all in the light inks. In fact some media won't handle it.
> > I recall a thread some time ago, Roy and a few others were talking
> > about the difficulty of partitioning and linearizing multiple K inks
> > along with variable dot. I think that's one reason the K7s are best
> > used with a 2880 set up. It's only one dot size, the smallest, making
> > so many inks easier to deal with. Also, with so many inks dithering in
> > and out, I don't see why one would even need variable dot.
> > For those less masochistic than I, the standard setup without variable
> > dots works very well, as you know, and the old versions do that just
> > fine. One step beyond that in v12 would be to set the K to open to the
> > large dot size, and maybe the lightest ink start with the smallest,
> > and gain some advantage there from the new feature. Other than
> > relinearizing, not much tweaking there.
> >
> > > Any idea how your tweaked quad output would compare to
non-tweaked hex-
> > > or sep-tone output in terms of print quality? In other words, if
I have
> > > a hex-tone set available, should I use that rather than going
through
> > > the pain and effort of tweaking the RIP for quads?
> >
> > Though I have buried my nose in some K7 output and am very impressed,
> > I have none here now for direct comparison. I don't intend to switch
> > to it so having it around taunting me would not be healthy.
> > I can say this is a step up for quads, subtle though. I sent the
> > tweaked dot setup to a very experienced pro friend, and he said he
> > made the best quad print he's ever made with it. Would most people see
> > that difference, or place a value on it? I think there is no beating
> > K7 output for neutral smooth B&W artifact free digital printing on art
> > papers that I have seen. I would not expect v12 quads with tweaked
> > dots to hold up to K7 under a loup. Without a loup, I don't know. I
> > need variable hues, so other considerations come to play.
> > I needed v12 for 9800 color, so had to get it anyway. Basically, I'm
> > reporting on miniscul obscure issues of interest to me. I hope it
> > didn't sound like an ad for the new version.
> > Both Roy and John Dean have working K6 setups right now, I think they
> > could say more about that quality.
> > Too bad we are all so far aprt, bringing everyone together with their
> > print and problems would be very informative.
> > Tyler
>