Digital BW, The Print

----- Original Message -----

From: "Roy Harrington" <roy@...>
To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
Sent: Monday, March 10, 2003 8:55 AM
Subject: [Digital BW] Ink Set Ideas and Questions (7500)


>
> All these new inks and ink sets have been very interesting.
> I'm finally going to take possession of a 7500 this week, so
> I have to order some inks.  I've liked the VM-Sepia inks for
> my 1160 and just getting those for the 7500 is certainly an
> easy solution.
>
> But, with 6 inks now and my QuadTone workflow I've got
> more control and options than an RGB workflow.
>
> With the CMYK workflow I've gotten the best dMax by using
> 100% black ink plus 100% cyan (dark gray) -- i.e. I can easily
> specify 200% ink usage for dMax.   So my idea is to try for
> using 2 inks that are black to get 200% pure black.  So the
> question is what ink to give up.

Interesting message Roy.

There are however few inkjet coatings that allow 200 % ink
without losing detail. Even with a linearised 9000 I usually get
no further than 175 % on Hahnemuhle coatings. Wasatch SoftRip,
the primary RGB colours mixed from CM MY YC. There's another
reason not to go any further, with pigment inks you will not gain
extra black nor in the case of colours extra gamut. The pigments
are just too opaque, for colour where no light gets through to
the substrate
so you do not get filtered light reflected, for black where no
light is lost in the depth of a dye pool it is reflected from the
surface of the pigment particles. Paul wrote a message just
before yours where he compared a double printed black with a
single one and not much difference between them. At the colorsync
list there was a short thread on the gamut gain with pigment inks
and where the ink limits should be set before making the profile
as for example a 100% cyan + 100% magenta doesn't deliver more
gamut than a slightly lower percentages.

There's a difference between Paul's double printing and printing
200 % ink on the paper in one go. Inkbleed in the depth and width
of the coating is increased with the last compared to the first
where there's drying in between. The density will probably be
higher with the 200% ink in one go.

There is however a good reason to add some dark grey (and in my
case composite grey) to even the darkest black. With the 9000 (
and it will not be much different with the other models of that
generation) you often get white lines in the black with pigment
inks. That is gone with the addition of a low percentage of grey
ink up to 000 RGB. Under Colour Addition so to speak. Before you
start with less grey inks I would recommend to print some samples
of black with details and check what is gained and what is lost
in terms of density, smoothness, detail. For smoothness it may
even be better to add small percentages of two greys into the
black. If you really go for two blacks then I think you will not
use 200% at all. Consider a gloss black and a matt black then,
there are other advantages than density with that set.

Ernst

Roy,

>...I'm finally going to take possession of a 7500 ...

Congratulations.

>... I've liked the VM-Sepia inks for
>my 1160 and just getting those for the 7500 is certainly an
>easy solution.

Yes, the vm-s is what I've used in my 3000 for some time.  I now have the
new VM 4.3 (.5 -- just a slightly modified toner) in the 7500, mostly
because it's the latest, but also because it is substantially more
lightfast.  However, the warm end doesn't reach the full-on sepia of the
vm-s.  The light sepia tone I used most with the vm-s inkset is about 0.20
units warm (Y - C = 0.20), whereas the warm end of the 4.3 inkset is about
0.10 units warm.  That is much closer to a light sepia than the old vm
inkset warm, but the jury is still out on whether it'll cut the competition
in mix media competitions.

I'm still considering a few options for a new vm-s.

>With the CMYK workflow ... my idea is to try for
>using 2 inks that are black to get 200% pure black.  So the
>question is what ink to give up.

>...

>Idea 3
>K - Black
>C - Black,   c - Dark Gray
>M - Medium Gray, m - Light Gray
>Y - Sepia

This is close to the vm-s/dual-K inkset I loaded into my 7500 at one point.
I used Epson Archival black in the cyan spot.  I wanted to have the ability
to print with a matte black but also the option to turn off that black with
software and use the RC-compatible black ink on RC papers.

The basic system worked and was even controllable with the Epson driver
using RGB curves in Photoshop.  However, I ran into several problems.
First, a strictly ink-related problem was that the Epson Archival inkset was
not the best -- and not close when diluted with the bases that were
available to me.  So, the inkset faded and warm-shifted worse than the
existing vm-s inkset.

I also had banding problems.  This could have been a problem with the
viscosity of the inks.  It could also be that the 7500 is more prone to
banding with radical RGB separation curves than are more modern
printers/drivers.

Finally, I found the dither to be too rough in spots.  This, again, could
have just been an interaction between the separation curves and the 7500
driver that your workflow might be able to avoid.  I think the dither
problems I ran into were related to the driver's built-in cross-overs of the
cyan/light cyan and magenta/light magenta.

>I can't control the Dark/Light need of Cyan and Magenta
>so those pairs have to the tied together.

Too bad.  That is one thing I'd really like to see in a RIP.

>So my question is what the dark/light differentiation should
>be for C-c and M-m and how to best  use that to get the best
>gray transitions.

My experience is that much more trouble results from having the light ink
too close to the full-strength than the other way around.

For the vm system, consistency with the existing inkset so that the existing
curves would work was the controlling issue.  When the original vm inks were
mixed, the spacing was based on the PiezoBW variable-tone that I'd started
with.  The M (light gray) was a compromise between the Piezo Y and M.  No
consideration was made of internal cross-overs, since I was dealing with an
1160.

When the vm inkset was made into a hextone inkset, the decision was also
rather pragmatic.  The existing MIS "45%" relative to the Piezo cyan was
close to the C/light-c difference in color inksets.  So, it was used and
worked.  For the toner, a simple 50% dilution worked well, so it was used.
I still use that for the toner, and it works well.

Some inksets use a 1/3 dilution for both C/c and M/m.  However, since the
ink dilution-density curves are far from linear, this set ratio for mixing
light inks is rather random.  On the other hand, it seems to work

For the vm-s/dual-K inkset I mixed for the 7500 I studied the issue more
closely.  I'd say the low end of the separation between the full-strength
and light inks that I've seen is the Photo K-Light K situation.  This is
about the same as the matte K (UC, Museum, or Eboni) on matte paper and the
Epson Archival K on matte paper.  So, I used that in my experimental dual-K
inkset, and it worked fine.

What I did was take the log density scale and simply spread my ink densities
evenly from the C (Epson Archival black-darkest "gray") to the
light-m/lightest gray, which I pegged at the Piezo Y density.

With the Epson Archival K as the cyan, the light-c would then be just about
the same as the existing hextone vm light-c (about 1.10 on EAM in the K
position of an 1160 and using BO printing).  This is about the same spread
as the PiezoBW K to C spread.  M would be about .75, and light-m about .40
(the Piezo/FS-Y density).

Of course, I also had a dither issue that might have been cross-over
related.  Perhaps if I'd had the C-c & M-m spread less, the
cross-over/dither problem may have been less.

>... I'm leaning toward the 3rd
>choice to get the smoothest grayscale and I'm hoping as good
>a sepia.

That's what I decided to do also.  The 3000 (even larger dots than the 7500)
and the vm-s with its singe toner looks rather good to me.  So, I went with
one toner in the Y spot.  I reduced the gamut of the vm-s toner, since I
found the light sepia curve was what I preferred.

> But do the grays go CcMm or CMcm?

Of course, to stay consistent with the Epson driver, I had to have the grays
in the order C (Arc/Photo K), c, M, m.

> Seems like between the FS-N and the VMS inks there ought
>to be many possibilities.

For the even-distribution approach I too, about the only ink that is not off
the shelf in the above system is the 0.75 density M.  The FS-N M should be
about .64.  So, it would need to be mixed with, perhaps 1/3 Hex light-c FS-N
to get to the .75 target.

On the other hand, if the dither from the cross-over was my problem, maybe
I'd been better off with C = Arc K, c = FSn=C, M = FSn-M, m = FSn-Y.  With
the 3000 we can cover the Piezo M to C spread with little trouble, so it
should also be doable with the 7500.  I was hoping to close that spread a
bit with the even distribution approach, but that could have been a mistake.

Have fun.  I'm waiting "patiently" (not) for a good RIP for my PC system.
Then I'll take another shot as a vm-s/dual-K system.

Paul
http://www.PaulRoark.com

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Ernst 
Dinkla" <E.Dinkla@c...> wrote:
> 
> ----- Original Message -----
> From: "Roy Harrington" <roy@h...>
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Sent: Monday, March 10, 2003 8:55 AM
> Subject: [Digital BW] Ink Set Ideas and Questions (7500)
> 
> 
> >
> > All these new inks and ink sets have been very interesting.
> > I'm finally going to take possession of a 7500 this week, so
> > I have to order some inks.  I've liked the VM-Sepia inks for
> > my 1160 and just getting those for the 7500 is certainly an
> > easy solution.
> >
> > But, with 6 inks now and my QuadTone workflow I've got
> > more control and options than an RGB workflow.
> >
> > With the CMYK workflow I've gotten the best dMax by using
> > 100% black ink plus 100% cyan (dark gray) -- i.e. I can easily
> > specify 200% ink usage for dMax.   So my idea is to try for
> > using 2 inks that are black to get 200% pure black.  So the
> > question is what ink to give up.
> 
> Interesting message Roy.
> 
> There are however few inkjet coatings that allow 200 % ink
> without losing detail. Even with a linearised 9000 I usually get
> no further than 175 % on Hahnemuhle coatings. Wasatch 
SoftRip,
> the primary RGB colours mixed from CM MY YC. There's 
another
> reason not to go any further, with pigment inks you will not gain
> extra black nor in the case of colours extra gamut. The 
pigments
> are just too opaque, for colour where no light gets through to
> the substrate
> so you do not get filtered light reflected, for black where no
> light is lost in the depth of a dye pool it is reflected from the
> surface of the pigment particles. Paul wrote a message just
> before yours where he compared a double printed black with a
> single one and not much difference between them. At the 
colorsync
> list there was a short thread on the gamut gain with pigment 
inks
> and where the ink limits should be set before making the 
profile
> as for example a 100% cyan + 100% magenta doesn't deliver 
more
> gamut than a slightly lower percentages.
> 
> There's a difference between Paul's double printing and 
printing
> 200 % ink on the paper in one go. Inkbleed in the depth and 
width
> of the coating is increased with the last compared to the first
> where there's drying in between. The density will probably be
> higher with the 200% ink in one go.
> 
> There is however a good reason to add some dark grey (and in 
my
> case composite grey) to even the darkest black. With the 9000 
(
> and it will not be much different with the other models of that
> generation) you often get white lines in the black with pigment
> inks. That is gone with the addition of a low percentage of grey
> ink up to 000 RGB. Under Colour Addition so to speak. Before 
you
> start with less grey inks I would recommend to print some 
samples
> of black with details and check what is gained and what is lost
> in terms of density, smoothness, detail. For smoothness it 
may
> even be better to add small percentages of two greys into the
> black. If you really go for two blacks then I think you will not
> use 200% at all. Consider a gloss black and a matt black then,
> there are other advantages than density with that set.
> 
> Ernst

Thanks, Ernst.

One of the things that I'm not too sure about is "what is 100%
of an ink".   With all the different resolutions and variable dot
sizes,  it seems that 100% is arbitrarily defined as a maximum
amount of ink that maintains some linearity from 0% to 100%.
This is just from testing with an 1160 using epson drivers and
now using gimp-print drivers -- I don't have an inside scoop
on the matter.

I've just experimentally found with the gimp-print driver that  if
I output the maximum of 2 inks Black and Dark Gray (cyan)
I get a greater dMax.  So I'm guessing it would be even
greater if I output the maximum of 2 Black inks.  Certainly if
I try this and dMax tops off before 200% I can easily cap it
where ever is best.

Roy

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul 
Roark" <paul.roark@v...> wrote:

LOTS of great info, Paul, thanks.

> 
> >I can't control the Dark/Light need of Cyan and Magenta
> >so those pairs have to the tied together.
> 
> Too bad.  That is one thing I'd really like to see in a RIP.

I think this is in works but I don't know when.

----------------------------------------


> 
> > But do the grays go CcMm or CMcm?
> 
> Of course, to stay consistent with the Epson driver, I had to 
have the grays
> in the order C (Arc/Photo K), c, M, m.

So the dark/light inks should be closer in density,  i.e.
better than interleaving.

> 
> > Seems like between the FS-N and the VMS inks there ought
> >to be many possibilities.
> 
> For the even-distribution approach I too, about the only ink that 
is not off
> the shelf in the above system is the 0.75 density M.  The FS-N 
M should be
> about .64.  So, it would need to be mixed with, perhaps 1/3 Hex 
light-c FS-N
> to get to the .75 target.

I'm inclined to go this way for the immediate time frame.
Do you offhand have densities and mixing proportions for
the hextone FS-N and VM-Sepia inks?  I can probably 
extrapolate within that bunch to come up with some trials.


> 
> Have fun.  I'm waiting "patiently" (not) for a good RIP for my PC 
system.
> Then I'll take another shot as a vm-s/dual-K system.

If you get any inkling to take an old PC, load it up with Linux,
and run my quadtone stuff, let me know.   With my Macs
I'm setting up an older G3 laptop as my print server -- its
a great way to off load all the RIP processing and have the
printer remotely connected by wireless or ethernet.

> 
> Paul
> http://www.PaulRoark.com

Thanks, again.
Roy