Digital BW, The Print

Hi Folks --

Please forgive me if this question has been done to death, but I'm 
looking at using Nevins' Transfer Curve tutorial 
(http://www.jnevins.com/transferfunctioncurve.htm) with Picture 
Window in order to calibrate my Epson 1520 with MIS quadtones and a 
variety of papers.  Can PW emulate the "Transfer Functions" dialog 
box shown in the tutorial?  Has anyone else had success in using this 
method with quadtone inks?  Thanks!!

Darrell

Darrell:

Try Transforms|Gray|Brightness Curve.  I recommend you set 19 points 
(plus the 2 end points) using Shift-click.  You can then print out a 
stepwedge and scan the results to determine if each point should be 
made lighter or darker.

A generic proceedure to create these curves (including RGB 
partitioned curves) can be found in Profile_make2.pdf in 
Files > Ink Sets, reviews and techniques > MIS FS.

Good Luck.

Jeff Randall


--- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
<darrelleifert@y...> wrote:
> Hi Folks --
> 
> Please forgive me if this question has been done to death, but I'm 
> looking at using Nevins' Transfer Curve tutorial 
> (http://www.jnevins.com/transferfunctioncurve.htm) with Picture 
> Window in order to calibrate my Epson 1520 with MIS quadtones and a 
> variety of papers.  Can PW emulate the "Transfer Functions" dialog 
> box shown in the tutorial?  Has anyone else had success in using 
this 

> method with quadtone inks?  Thanks!!
> 
> Darrell

Jeff --

Thanks for the pointer.  I'll try and adapt your workflow to my Epson 
2450 and see how far I can get.  I've used Paul Roark's .AMP files 
with PW but I'm getting some muddy midtones on my 1520, and so would 
like to "start from scratch" (alas, no $ for Cone's Piezography 
driver).

Cheers,
Darrell

--- In DigitalBlackandWhiteThePrint@y..., "jrandall1149" 
<jrandall@c...> wrote:
> Darrell:
> 
> Try Transforms|Gray|Brightness Curve.  I recommend you set 19 
points 
> (plus the 2 end points) using Shift-click.  You can then print out 
a 
> stepwedge and scan the results to determine if each point should be 
> made lighter or darker.
> 
> A generic proceedure to create these curves (including RGB 
> partitioned curves) can be found in Profile_make2.pdf in 
> Files > Ink Sets, reviews and techniques > MIS FS.
> 
> Good Luck.
> 
> Jeff Randall
> 
> 
> --- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
> <darrelleifert@y...> wrote:
> > Hi Folks --
> > 
> > Please forgive me if this question has been done to death, but 
I'm 
> > looking at using Nevins' Transfer Curve tutorial 
> > (http://www.jnevins.com/transferfunctioncurve.htm) with Picture 
> > Window in order to calibrate my Epson 1520 with MIS quadtones and 
a 
> > variety of papers.  Can PW emulate the "Transfer Functions" 
dialog 

> > box shown in the tutorial?  Has anyone else had success in using 
> this 
> > method with quadtone inks?  Thanks!!
> > 
> > Darrell

Darrell:

I assumed that you wanted to create a lumped RGB curve (a non-
partitioned workflow that combines the three gray inks into one thus 
yielding a 2 toned ink set -- gray and black).  For this kind of 
workflow, only one curve is needed.  Examples are Nivin's and Woolf's 
workflows.  The workflow I used to generate the curves for my Epson 
1160 is a partitioned RGB workflow where a seperate "transfer" curve 
is created for each of the three gray inks. It, however, can easily 
be adapted to creating a single lumped RGB curve.  I have sucessfully 
used it to create a lumped RGB curve for my 1160 starting with 
Woolf's curve. 

Remember, that although you can not *directly* enter the I/O values 
of curves in Picture Windows, you can open the .crv (grayscale) 
and .cc (color) files in a text editor and tweak the values.

I recommend that you start your journey with developing a lumped 
workflow to gain experience.  Also lay in a large supply of ink and 
paper and note that the cost in time and supplies may approach the 
cost of the barebones Piezo system...  (;>)) 

Contact me here or offline if you have additional questions.

Jeff Randall


--- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
<darrelleifert@y...> wrote:
> Jeff --
> 
> Thanks for the pointer.  I'll try and adapt your workflow to my 
Epson 
> 2450 and see how far I can get.  I've used Paul Roark's .AMP files 
> with PW but I'm getting some muddy midtones on my 1520, and so 
would 
> like to "start from scratch" (alas, no $ for Cone's Piezography 
> driver).
> 
> Cheers,
> Darrell
> 
> --- In DigitalBlackandWhiteThePrint@y..., "jrandall1149" 
> <jrandall@c...> wrote:
> > Darrell:
> > 
> > Try Transforms|Gray|Brightness Curve.  I recommend you set 19 
> points 
> > (plus the 2 end points) using Shift-click.  You can then print 
out 
> a 
> > stepwedge and scan the results to determine if each point should 
be 
> > made lighter or darker.
> > 
> > A generic proceedure to create these curves (including RGB 
> > partitioned curves) can be found in Profile_make2.pdf in 
> > Files > Ink Sets, reviews and techniques > MIS FS.
> > 
> > Good Luck.
> > 
> > Jeff Randall
> > 
> > 
> > --- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
> > <darrelleifert@y...> wrote:
> > > Hi Folks --
> > > 
> > > Please forgive me if this question has been done to death, but 
> I'm 
> > > looking at using Nevins' Transfer Curve tutorial 
> > > (http://www.jnevins.com/transferfunctioncurve.htm) with Picture 
> > > Window in order to calibrate my Epson 1520 with MIS quadtones 
and 
> a 
> > > variety of papers.  Can PW emulate the "Transfer Functions" 
> dialog 
> > > box shown in the tutorial?  Has anyone else had success in 
using 

> > this 
> > > method with quadtone inks?  Thanks!!
> > > 
> > > Darrell

Darrell:

I forgot to mention that I recommend that you develop your workflow 
with the Epson driver set to No Color Correction (and heavy weight 
matte, 1440dpi, highest quality dither, highspeed off. Also use 
AdobeRGB 1998 if developing a partitioned workflow).  Paul Roark's 
latest VM partioned workflows take this approach as well as my FS 
workflow.  It greatly simplifies life, imo.

Jeff Randall

--- In DigitalBlackandWhiteThePrint@y..., "jrandall1149" 
<jrandall@c...> wrote:
> Darrell:
> 
> I assumed that you wanted to create a lumped RGB curve (a non-
> partitioned workflow that combines the three gray inks into one 
thus 
> yielding a 2 toned ink set -- gray and black).  For this kind of 
> workflow, only one curve is needed.  Examples are Nivin's and 
Woolf's 
> workflows.  The workflow I used to generate the curves for my Epson 
> 1160 is a partitioned RGB workflow where a seperate "transfer" 
curve 
> is created for each of the three gray inks. It, however, can easily 
> be adapted to creating a single lumped RGB curve.  I have 
sucessfully 
> used it to create a lumped RGB curve for my 1160 starting with 
> Woolf's curve. 
> 
> Remember, that although you can not *directly* enter the I/O values 
> of curves in Picture Windows, you can open the .crv (grayscale) 
> and .cc (color) files in a text editor and tweak the values.
> 
> I recommend that you start your journey with developing a lumped 
> workflow to gain experience.  Also lay in a large supply of ink and 
> paper and note that the cost in time and supplies may approach the 
> cost of the barebones Piezo system...  (;>)) 
> 
> Contact me here or offline if you have additional questions.
> 
> Jeff Randall
> 
> 
> --- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
> <darrelleifert@y...> wrote:
> > Jeff --
> > 
> > Thanks for the pointer.  I'll try and adapt your workflow to my 
> Epson 
> > 2450 and see how far I can get.  I've used Paul Roark's .AMP 
files 
> > with PW but I'm getting some muddy midtones on my 1520, and so 
> would 
> > like to "start from scratch" (alas, no $ for Cone's Piezography 
> > driver).
> > 
> > Cheers,
> > Darrell
> > 
> > --- In DigitalBlackandWhiteThePrint@y..., "jrandall1149" 
> > <jrandall@c...> wrote:
> > > Darrell:
> > > 
> > > Try Transforms|Gray|Brightness Curve.  I recommend you set 19 
> > points 
> > > (plus the 2 end points) using Shift-click.  You can then print 
> out 
> > a 
> > > stepwedge and scan the results to determine if each point 
should 
> be 
> > > made lighter or darker.
> > > 
> > > A generic proceedure to create these curves (including RGB 
> > > partitioned curves) can be found in Profile_make2.pdf in 
> > > Files > Ink Sets, reviews and techniques > MIS FS.
> > > 
> > > Good Luck.
> > > 
> > > Jeff Randall
> > > 
> > > 
> > > --- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
> > > <darrelleifert@y...> wrote:
> > > > Hi Folks --
> > > > 
> > > > Please forgive me if this question has been done to death, 
but 
> > I'm 
> > > > looking at using Nevins' Transfer Curve tutorial 
> > > > (http://www.jnevins.com/transferfunctioncurve.htm) with 
Picture 

> > > > Window in order to calibrate my Epson 1520 with MIS quadtones 
> and 
> > a 
> > > > variety of papers.  Can PW emulate the "Transfer Functions" 
> > dialog 
> > > > box shown in the tutorial?  Has anyone else had success in 
> using 
> > > this 
> > > > method with quadtone inks?  Thanks!!
> > > > 
> > > > Darrell

--- In DigitalBlackandWhiteThePrint@y..., "jrandall1149" 
<jrandall@c...> wrote:> 
--snip
The workflow I used to generate the curves for my Epson 
> 1160 is a partitioned RGB workflow where a seperate "transfer" 
curve 
> is created for each of the three gray inks.
--snip 

Jeff --

Hmmm, the light is slowly dawning.  When I apply a "canned" .AMP 
transfer curve to a 21-step grayscale wedge converted to color in 
Picture Window, the stepwedge visually breaks down into red / 
yellow / cyan bands, each band flowing from dark to light.  I assume 
these are the separate transfer curves designed to partition the 
image into dark / midtone / light sections thereby using the gray 
inks in the MIS cartridge without the dot patterns caused by black 
ink halftoning.  OK, clear enough.

Now the "newbie" questions:

1)  How is what is essentially a CMY breakdown initially created 
within an RGB curves workspace?

2)  When the stepwedge with CMY transfer curves applied is printed 
and, say, the 70% through 85% band and the 40% through 50% bands are 
a solid gray, how is (are) the RGB curve(s) tweaked so that the solid 
blocks resolve into distinct 5% bands?  Can the stepwedge be printed, 
scanned, and some kind of histogram readout give a clue as to how the 
RGB curves can be adjusted to achieve the 5% separation?  

Put another way, is there a feedback loop that will provide 
corrective information on shaping the correct partitioned curves for 
each printer / paper combination, or is it massive trial and error?

Thanks again for your time.

Cheers,
Darrell

--- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
<darrelleifert@y...> wrote:
<snip>

> Now the "newbie" questions:
> 
> 1)  How is what is essentially a CMY breakdown initially created 
> within an RGB curves workspace?

Each RGB curve controls the output of one of the gray inks.  Here are 
a couple of paragraphs from the MIS-FS Partitioned Quad workflow 
document in the Files section of this forum:  

Files > Ink Sets, reviews and techniques > MIS FS > 1160 Partitioned 
Workflow  

What is a Partitioned Quad Workflow?
An RGB curve set is applied to the image and partitions (or assigns) 
each of the four inks to applicable overlapping ranges of grayscale 
values.  Thus the lightest ink prints the lightest portions of the 
gray scale etc. yielding essentially dotless lighter tones using the 
Epson driver.  Examples of partitioned workflows include those 
developed by Paul Roark, Chris Brandon, and others.   Lumped 
workflows, such as John Woolf's or Jerry Nivins' where the C, M, Y 
position inks are lumped or mixed together in approximate equal 
proportions across the grayscale range, yield essentially a two-tone 
ink system –- black and gray.  These lumped workflows can produce 
fine quality prints, however they are not dotless in the lighter 
tones (especially noticeable under a loupe) and sometimes show 
banding in the shadow areas.

Ink Positions
The Full Spectrum inkset uses the same ink positions as the 
PiezographyBW inkset.  Each FS ink bottle is labeled "Cyan" 
(C),  "Magenta" (M), "Yellow" (Y), or Black (K).  The C-ink is dark 
gray, the M-ink is medium gray, and the Y-ink is light gray.   As a 
labeling QC check, you may wish to smear some ink from each bottle 
onto a white sheet of paper with a q-tip and let dry.  If using a CFS 
or CIS system, make sure to connect the system tubes to the correct 
ink bottle.  The tubes running to the installed "color" cartridge 
should deliver  C, M, Y ink from left to right (as you stand in front 
of the printer). The K-ink tube goes to the black cartridge.  

Note:  The single curve of a lumped workflows controls all three of 
the gray inks--hense "lumped" or "nonpartitioned".  The black ink in 
both kinds of workflow is controlled directly by the Epson driver.  
The black ink is turned on when all three other inks are greater than 
about 50%.

> 2)  When the stepwedge with CMY transfer curves applied is printed 
> and, say, the 70% through 85% band and the 40% through 50% bands   
> are a solid gray, how is (are) the RGB curve(s) tweaked so that the 
> solid blocks resolve into distinct 5% bands?  

I am not sure I understand your question, but here goes... The 21 
steps (5% increments) of the stepwedge represent 21 points on a 
continious curve connecting the dots (I hope the FBI and CIA aren't 
reading (;>)) ).  If you printed a grayscale ramp, theoretically all 
shades of gray should be printed. Your example would represent 2 
bands of the continious grayscale ramp and would create a histogram 
that shows two large steps 40 to 50% and 70 to 85%. They would not 
resolve into 5% bands. An example of what a 21 stepwedge histogram 
looks like is shown in the profiles_make2.pdf document referenced 
eariler in this thread.

> Can the stepwedge be printed, 
> scanned, and some kind of histogram readout give a clue as to how  
> the RGB curves can be adjusted to achieve the 5% separation?  
> Put another way, is there a feedback loop that will provide 
> corrective information on shaping the correct partitioned curves   
> for each printer / paper combination, or is it massive trial and   
> error?

This process of scanning, using a histogram (Picture Windows Curves) 
or a spectorphotometer/densiometer to evaluate the curves, tweaking 
the curves, and reprinting is described in the profiles_make2.pdf 
document located in:

Files > Ink Sets, reviews and techniques > MIS FS.

Unfortunately there typically is the need to make quite a few 
iterations.


Good Luck.

Jeff Randall

Jeff --

Thanks for the detailed response.  I'll do some careful reading in 
the Files section.  To clarify question (2), when I print a stepwedge 
using a TS curve (specifically 4t_eam.amp) that shows an even CMY 
breakdown on screen, the printed version shows a sharp tonal break 
between 90% and 85%, and two groups of three bands (100, 95, 90 and 
85, 80, 75) print as a dark solid tone, with no discernable 
difference.  This produces landscapes for example, where parts of a 
grassy field suddenly go "flat" because certain defined tones in the 
negative are indistinguishable from each other in the print (obvious 
problem).  On the other hand, bands 5% through 70% show an evenly 
spaced progression.  I'm assuming (hoping) that some kind of TS curve 
adjustment can fix the problem.

Cheers,
Darrell

Darrell:

Several comments:

1)  The apparent evenness of the false colors (CMY) applied to your 
image is in no real way indicative of an even printed grayscale.

2)  Chris Brandon's workflow (eg., curve 4t_eam.amp, et.al.) is 
specifically set up for MIS Original inks.  These inks have different 
densities than the MIS FS inkset and therefore would not initially 
work well without significant modification.  *MOREOVER*, the 
positions of the inks in the MIS-Orig inkset are Red/Cyan=light gray, 
Green/Magenta=dark gray, and Blue/Yellow=medium gray.  The MIS-FS 
inkset (and PiezoBW inkset) positions are Red/Cyan=dark gray, 
Green/Magenta=midium gray, and Blue/Yellow=light gray.  

You didn't say which inkset you were using, but position *DOES* 
matter.  Of course you can locate your three gray inks in any 
positions and create curves for them that will work well for you, 
but....

3)  The profile_make2.pdf document broadly outlines the proceedure I 
use to modify curves.  More specifically for instance, if the 10% 
step (light end) is too dark, I would move the 10% control point on 
the Blue curve up a little (this discussion assumes the MIS-FS ink 
positions)  If the 85% step (dark end) is too light, I'd move the 85% 
control point on the Red curve down a little.  For steps between 
about 30% and 60%, control point moves on one or more of the 
individual R,G,B curves might be applicable. Note:  The 30-60% region 
is where the individual gray inks overlap.

4)  I again recommend you gain experience modifying single curve 
workflows (i.e., lumped RGB workflow curves like Woolf's and 
Nivin's). I also recommend you turnoff the Epson driver color 
control.  Position doesn't seem to matter with these lumped workflows.

Learn to walk before you fly.

Happy Trails.

Jeff Randall



--- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
<darrelleifert@y...> wrote:
> Jeff --
> 
> Thanks for the detailed response.  I'll do some careful reading in 
> the Files section.  To clarify question (2), when I print a 
stepwedge 
> using a TS curve (specifically 4t_eam.amp) that shows an even CMY 
> breakdown on screen, the printed version shows a sharp tonal break 
> between 90% and 85%, and two groups of three bands (100, 95, 90 and 
> 85, 80, 75) print as a dark solid tone, with no discernable 
> difference.  This produces landscapes for example, where parts of a 
> grassy field suddenly go "flat" because certain defined tones in 
the 
> negative are indistinguishable from each other in the print 
(obvious 
> problem).  On the other hand, bands 5% through 70% show an evenly 
> spaced progression.  I'm assuming (hoping) that some kind of TS 
curve 

> adjustment can fix the problem.
> 
> Cheers,
> Darrell

Hi Jeff --

I think I'm still a bit confused on several points (BTW, I am using 
the MIS original inksets):
> 
> 1)  The apparent evenness of the false colors (CMY) applied to your 
> image is in no real way indicative of an even printed grayscale.

I understand that the CMY curves break up the print so that dark, 
medium and light tones are printed by the appropriate cartridges.  
But if a grayscale stepwedge does not print evenly, are changes best 
made to the CMY color curves (when the stepwedge is converted to 
color) or a single 19-step brightness curve when the image is in 
grayscale mode?

--snip
>the 
> positions of the inks in the MIS-Orig inkset are Red/Cyan=light 
gray, 
> Green/Magenta=dark gray, and Blue/Yellow=medium gray.

If the gray inks (25%, 50% and 75%) are keyed to CMY color values, 
how can Red/Cyan, Green/Magenta and Blue/Yellow refer to the same 
tone/tank?  Logically it seems that a "red" curve would use half 
magenta and half yellow, meaning that a mix of the 75% and 50% tanks 
would print, not the exclusive 25% from a true Cyan curve. (I'm 
really confused on this one.)  

--snip
> 4)  I again recommend you gain experience modifying single curve 
> workflows (i.e., lumped RGB workflow curves like Woolf's and 
> Nivin's). I also recommend you turnoff the Epson driver color 
> control.  Position doesn't seem to matter with these lumped 
workflows.
> 
> Learn to walk before you fly.
> 

<sigh>  I would like to work with a single curve workflow, but doing 
so with my older Epson 1520 produces very noticeble dots in the light 
areas of the print. On the other hand, applying Chris Brandon's 
4t_eam.amp curve gave me very smooth, dotless light grays 
but "clumped" darks and midtones.  I hope therefore that if I can 
understand how he produced his original curves for the 1160, I can 
modify them to work properly with my 1520.  

Again, thanks for all your help -- the profile_make2.pdf did clarify 
many of my original questions.

Cheers,
Darrell

Darrell:

--- In DigitalBlackandWhiteThePrint@y..., "darrelleifert" 
<darrelleifert@y...> wrote:

> I think I'm still a bit confused on several points (BTW, I am using 
> the MIS original inksets):

Aaah.  Just like your doctor, lawyer, and accountant, you need to 
divulge all the key facts to focus the advice you get (;>))

> > 1)  The apparent evenness of the false colors (CMY) applied to
> > your image is in no real way indicative of an even printed
> > grayscale.

> I understand that the CMY curves break up the print so that dark, 
> medium and light tones are printed by the appropriate cartridges.  
> But if a grayscale stepwedge does not print evenly, are changes
> best made to the CMY color curves (when the stepwedge is converted
> to color) or a single 19-step brightness curve when the image is in 
> grayscale mode?

First, its the RGB curves that partition the image and trick the 
Epson driver into thinking it is printing various shades of CYMK.

Second, I am a firm believer in going to the source.  I'd fix the 
partitioning curves, especially if there are significant "flat 
spots".  Some of the advantages of partitioning are lost if you tweak 
the image too much with a grayscale transfer curve (which is 
essentially what Woolf's and Nivin's lumped workflows do).  Other 
folks may have a different opinion.

 --snip
> > The positions of the inks in the MIS-Orig inkset are
> > Red/Cyan=light gray, Green/Magenta=dark gray, and
> > Blue/Yellow=medium gray.

> If the gray inks (25%, 50% and 75%) are keyed to CMY color values, 
> how can Red/Cyan, Green/Magenta and Blue/Yellow refer to the same 
> tone/tank?  Logically it seems that a "red" curve would use half 
> magenta and half yellow, meaning that a mix of the 75% and 50%
> tanks would print, not the exclusive 25% from a true Cyan curve.
> (I'm  really confused on this one.)

I am outside my area of expertise, but here is my understanding.  
Maybe someone in this forum can jump in and add or give us links to 
color theory/print driver primers.

Your monitor (and editing software) works in RGB additive light space 
(red light + green light = yellow light).  Your printer lays down 
dyes or pigments that absorb part of the incident light and reflect 
the rest (cyan ink absorbs all the red light and reflect the rest). 
The Epson driver converts the RGB image into CYMK space which 
controls the CYMK ink colors in the printer carts.  

I think it is now the time to invoke "faith based printing" and just 
say that adjusting the red curve in a .amp, .cvr, or .cc curve set 
controls the amount of cyan that shows up on your grayscale 
(converted to 24-bit) image displayed on your monitor when the curve 
is applied which in turns controls the ink in the printers cyan 
position no matter what color or shade of gray may be there, etc.  

Also for clarification, the MIS Origial Quadtone inkset labels are 
NOT related to the actual densities -- K = 100%, "75" = 96%, "50" = 
79%, "25" = 24%.  Your milage (densities) may vary--however these 
values are ball park.  Only MIS knows why the ink labels and actual 
densities are this way.

 --snip
> > 4)  I again recommend you gain experience modifying single curve 
> > workflows (i.e., lumped RGB workflow curves like Woolf's and 
> > Nivin's). I also recommend you turnoff the Epson driver color 
> > control.  Position doesn't seem to matter with these lumped 
> > workflows.

> <sigh>  I would like to work with a single curve workflow, but
> doing so with my older Epson 1520 produces very noticeble dots in
> the light areas of the print. On the other hand, applying Chris
> Brandon's 4t_eam.amp curve gave me very smooth, dotless light grays 
> but "clumped" darks and midtones.  I hope therefore that if I can 
> understand how he produced his original curves for the 1160, I can 
> modify them to work properly with my 1520.  
> 
> The profile_make2.pdf did clarify many of my original questions.

You really don't need to understand HOW Chris developed his curves.  
All you really need to do is look at his .amp curves (that you have 
converted to PW .cc--right?) and follow the outline presented in 
profiles_make2.pdf ( Files> Ink Sets, reviews and techniques> MIS 
FS ). For example, if you have a flat spot at 70-80% and the 
histogram shows the center of the spot is at 75%, then you would move 
the 70% control point up (lighter) and the 80% control point down 
(darker), reprint and scan your stepwedge to check to results. This 
is an easy process using Picture Windows and its superior 
curves/histogram feature.  It however, is labor, ink and paper 
intensive, but NOT rocket science.

You may want to reduce the number of control points in each of Chris' 
curves to 21.  This again is very easy using PW.  Convert 
the PhotoShop .amp file to a PW .cc file and use a text editor to 
whack out the "extra" control points.

Hope this helps keep you on the path that ultimately leads to the 
ability to create art.  

Jeff Randall