Digital BW, The Print

Martin,

You wrote:

>MIS VM Initial Impressions and Questions

>...new Epson 1280 ...Perfect nozzle  checks ...

Could you tell us what the order of colors is with the 1280 nozzle check.
Once there are quad inks in there, it's not obvious to users which nozzle is
which.

>... the MIS VM inks so I assumed that the
>inks were to go into the standard positions and
>matched them to the labels on the CIS tubing.

Each bottle should be marked with the ink color position.

>... The curves were all dated 7/17/01.

I have updated curves files.

>... resampling before sharpening gave results that
>appear to be equally sharp [compared to Piezo].

I've run resolution tests comparing the Piezo and Epson drivers.  On my 1160
they are so close that the differences would not be visible.

>... my first questions.

>What is the preferred image print dpi for use with this system of
>printer and inks?

I've heard different stories with respect to this.  I think the answer would
be the same as with the Epson driver and any color inkset.

My practice is to print at 360 image dpi (almost always the top printer
resolution setting, although I often can't see any difference between 720
and 1440.

With respect to image file resolution, I found I could barely detect the
difference between 360 and 720.  I really don't know if anyone has good
enough eyes to see the differences.  I consider 360 so good that the hassle
of files larger than that way out-weighs any possible visual sharpness
difference.

I also often print at any resolution and let the driver resample.  However,
I've been told by people I respect that it is better to so the resampling in
Photoshop.  And when it's an image I really care about and the original scan
is not at a final print resolution of 360 I resample with Genuine Fractals.

>Where in the workflow do you resample to that resolution?

>Do you sharpen in grayscale or after the conversion to RGB or after
>the application of the curve?

I never convert to RGB and apply the curve until the image is all done
(except for split toning).  I want to save a final g/s image in g/s mode
just before I print.  I consider the conversion and curve application to be
nothing more than a printing function.

>Under magnification I see no dots, no pattern, no microscopic or "sub
>micro" banding, no window screen pattern.

That is as it should be with that printer.

>I tried a neutral print on Legion Photo Matte and I think that it is
>a better match than the EAM. ... Paul's curve seems to work well
>on both of them.

I just received my first box of Legion Photo Matte.  I'll do a fade test
soon.  (Museo is currently in the fade tester.)

>Blacks seem comparable to Piezo blacks.
>The MIS may appear a tiny bit darker ...

The MIS VM black ink is virtually (if not actually) the same as Piezo black
ink.  On my 1160, the Epson driver prints darker than the Piezo driver.

>... the medium-warm and warm curves, vmh-mw6
>and vmh-w9, on EAM. Both of these posterized heavily.

Let me send you the newer versions.  I probably ought to have MIS change
what is on their website also.

>... The worst was the warm curve with the magenta slider
>at \ufffd25 to see what would happen by pushing the system to
>maximum warm.

I've changed my approach for the warm curve.  Rather than have people do
this and get bad results, I'm now writing the warm curve to be as warm as
the system will go.  If people want to fine-tune warmth they should use the
medium warm curve.  The warm curve is written in such a manner that it just
does not respond well to the sliders.  So, the new warm curve (I'm going to
do this for all the printer types) is pre-set at maximum warmth.  Don't use
sliders with it.

>The warm curves gave a print tone that was similar to Piezo but does
>not appear to have the slight green cast I have referred to
>previously.

MIS is actually a little warmer than Piezo.  Piezo has a "warmth" (that I
define as red channel reading minus blue channel reading) of about 7 to 8,
MIS is at about 12 to 13.

I also tried the cool curve vmh-c13 and this also printed well.
Similar to the neutral curve but it is definitely bluer.


>One thing I did not find in the Advance area of the
>Epson driver was a "Half toning" setting which is
>supposed to be set to  "error diffusion".

I may have been informed that Epson dropped that name on the newest
printers.  It may indicate print quality so other way.

>If you have posterizing with a given curve and paper combination,
>what is the workflow for creating a correction curve to achieve the
>desired results? Assuming no instruments other than your eyes.

First, get the updated curves.

If the curve you used is final and works with the papers we know are
compatible, use the 21-step test file and either a scanner or eyes to match
the densities to a known output that looks good.  The Transfer Function in
the print driver can be used to profile a paper.  That way you don't have to
mess with the actual curves, which are much trickier to modify.

If a paper becomes very popular, I may re-write some of the curves to avoid
this.  I will also be writing some notes on the little tricks I've found
help for modifying curves.

I sure hope others will help do this profiling and share their
results/files.

>All in all I find the printer/ink combination very promising and I
>certainly like the color.

Great.  (And you even end up saving a bunch of money on ink, not to mention
no software expenses.)

Paul
http://www.PaulRoark.com