Digital BW, The Print

For those gurus who understand greyscale partitioning way better than I do I have the 
following question.  Let's say you printed QTR's ink separation page and determined that you 
would limit the K ink at 50% but noticed that your LK and LLK inks continued to show solid 
increases in density well above 50%.  One could either limit each ink at 50% and calculate the 
partitioning points based on this ink limit or decide to limit K at 50% but run LK/LLK at a 
higher ink limit and calculate partitioning points based on this mix of limits.  Is there 
anything to be gained by the latter?  I was thinking that perhaps one would ordinarily prefer 
to run inks close to their "peak" densities in order to get a less "dotty" greyscale. Does this 
make sense or does it not really matter?

Cheers

Steve

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Steve Kale" <stevekale@b...> wrote:
> ... I was thinking that perhaps one would ordinarily prefer 
> to run inks close to their "peak" densities in order to get a less "dotty" greyscale. Does this 
> make sense or does it not really matter?

Yes it does make sense. The most effective solution for how to get there with QTR, I don't 
know. I ran into the same sort of conundrum with it, but haven't gone back to play with it.
Tyler

Thanks Tyler.  What was the conundrum?  It is easy to implement in practice
if one thinks it will benefit.... (except that I banged into issues with
Roy's new linearization input check when trying to use 51 step data: if a
couple of the points come too close to each other the check calls an invalid
linearization data)...unless there is something that gets screwed up behind
the scenes when QTR calcs the partitioning.

> From: Tyler Boley <tyler@...>

> 
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Steve Kale"
> <stevekale@b...> wrote:
>> ... I was thinking that perhaps one would ordinarily prefer
>> to run inks close to their "peak" densities in order to get a less "dotty"
>> greyscale. Does this
>> make sense or does it not really matter?
> 
> Yes it does make sense. The most effective solution for how to get there with
> QTR, I don't 
> know. I ran into the same sort of conundrum with it, but haven't gone back to
> play with it.
> Tyler
>

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Steve Kale <stevekale@b...> wrote:
> Thanks Tyler.  What was the conundrum?

How to survive without sleep so I can go back and play with it and a thousand other things...

oh, and have a life.

Tyler

Ah yes well that I understand.  It's actually sunny in London for a change!

> From: Tyler Boley <tyler@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Wed, 13 Jul 2005 15:18:24 -0000
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: [Digital BW] Re: Partitioning greys - an ink limit question
> 
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, Steve Kale
> <stevekale@b...> wrote:
>> Thanks Tyler.  What was the conundrum?
> 
> How to survive without sleep so I can go back and play with it and a thousand
> other things...
> 
> oh, and have a life.
> 
> Tyler

If you can give me a little more info, I might be able to help. This 
same thing comes up with the color channels too, even more so when you 
start using odd inks like I do.

What I need to know is, does QTR linearize each channel, and then you 
make the mixing partitions (resulting in a final linear "color"). Or do 
you make the mixes and then linearize that resulting "color". Sorry I'm 
not as familiar with QTR as I maybe should be, but I use a different 
RIP.

Obviously you want you full black to run up to full density (or your 
target density), so there isn't much to do there.

QTR creates a greyscale from predetermined inks and then linearizes this
greyscale.  The process, in simple terms, is determine ink limits for each
component of the greyscale; print a calibration page which prints a
greyscale for each ink given the determined ink limits; measure the peak
densities of the lighter inks and determine where they sit on the K
greyscale, eg peak LK is 45% K, LLK is 15% K. QTR constructs a greyscale
using these inks, their limits and the partitioning points.  Print a
greyscale with this "ink curve".  Measure the L* and feed this info back
into the curve as a linearization function.  The maths behind the smoothing
of rolling from one ink to the next in the greyscale is hidden and well
beyond me.  So if one sets the ink limit of LK higher, all else being equal
it will print darker at peak and hence be 'partitioned' at a higher
percentage of K.

> From: dfaprinting <dfaprinting@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Wed, 13 Jul 2005 15:34:43 -0000
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: [Digital BW] Re: Partitioning greys - an ink limit question
> 
> If you can give me a little more info, I might be able to help. This
> same thing comes up with the color channels too, even more so when you
> start using odd inks like I do.
> 
> What I need to know is, does QTR linearize each channel, and then you
> make the mixing partitions (resulting in a final linear "color"). Or do
> you make the mixes and then linearize that resulting "color". Sorry I'm
> not as familiar with QTR as I maybe should be, but I use a different
> RIP.
> 
> Obviously you want you full black to run up to full density (or your
> target density), so there isn't much to do there.
>

Steve, I am using piezo carbon sepia inks and QTR and was just asking 
myself the same question as you.  I asked Tom Moore on the QTR forum 
and he said it didn't really matter.  My K maxes out at around 65% and 
the other inks at 85%.  He said to just use 65% and not sweat it.

I asked him about printing the ink separation page at 65% and then 
again at 85% and then comparing/measuring the densities of the lighter 
inks (off the 85% page) against the K on the 65% page.  When creating 
the curve, I was going to put the ink limit for the K at 65% and the 
others at 85% and then linearize.  Tom said I didn't really need to 
bother with all of that.

Jenny

> From: byushooter <jellerbe@...>

> 
> Steve, I am using piezo carbon sepia inks and QTR and was just asking
> myself the same question as you.  I asked Tom Moore on the QTR forum
> and he said it didn't really matter.  My K maxes out at around 65% and
> the other inks at 85%.  He said to just use 65% and not sweat it.
> 
> I asked him about printing the ink separation page at 65% and then
> again at 85% and then comparing/measuring the densities of the lighter
> inks (off the 85% page) against the K on the 65% page.

This is exactly the exercise I was going through.

> When creating 
> the curve, I was going to put the ink limit for the K at 65% and the
> others at 85% and then linearize.  Tom said I didn't really need to
> bother with all of that.
> 


I am sorry but I don't know who Tom Moore is. Sorry Tom. At any rate, this
is the question I am thinking about.  From a very basic level, it would make
sense to ensure maximum ink coverage throughout the scale and that would
imply running the lighter inks further up the scale (then you get your
'mid-tone' density through more ink rather than less).  Whether from a
practical point of view it makes a difference I don't know.  But if people
see benefit in more than 3 greys then this would imply that there is benefit
in at least using 3 fully.

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Steve Kale <stevekale@b...> 
wrote:
> 
> I am sorry but I don't know who Tom Moore is. Sorry Tom. At any rate, this
> is the question I am thinking about.  From a very basic level, it would make
> sense to ensure maximum ink coverage throughout the scale and that would
> imply running the lighter inks further up the scale (then you get your
> 'mid-tone' density through more ink rather than less).  Whether from a
> practical point of view it makes a difference I don't know.  But if people
> see benefit in more than 3 greys then this would imply that there is benefit
> in at least using 3 fully.

I believe Tom Moore wrote a tutorial for the QTRGui for Windows users.

So, if you're using different ink limits for different inks, say a limit of 60 for K and a limit 
of 80 for LC, how would you calculate the adjusted Gray Value?

Also, another question:  in the UT7 inkset, I find that LK is not being used in the QTR 
profiles.  Does that affect the Overlap percentage (which seems to be 16% in the profiles 
I've looked at)?  In other words, if you are using a gray lighter than LK for overlapping with 
K, might overlapping it decrease your DMax, rather than increasing it?

Ed

> From: edrudolpho <erudolph@...>

> 
> I believe Tom Moore wrote a tutorial for the QTRGui for Windows users.
> 

Ah I recall the name now


> So, if you're using different ink limits for different inks, say a limit of 60
> for K and a limit
> of 80 for LC, how would you calculate the adjusted Gray Value?

Print the calibration page at 60 then at 80.  Measure LC off the 80 print
and compare with K on the 60 print.

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Steve Kale 
<stevekale@b...> wrote:
> QTR creates a greyscale from predetermined inks and then linearizes 
this
> greyscale.  The process, in simple terms, is determine ink limits 
for each
> component of the greyscale; print a calibration page which prints a
> greyscale for each ink given the determined ink limits; measure the 
peak
> densities of the lighter inks and determine where they sit on the K
> greyscale, eg peak LK is 45% K, LLK is 15% K. QTR constructs a 
greyscale
> using these inks, their limits and the partitioning points.  Print a
> greyscale with this "ink curve".  Measure the L* and feed this info 
back
> into the curve as a linearization function.  The maths behind the 
smoothing
> of rolling from one ink to the next in the greyscale is hidden and 
well
> beyond me.  So if one sets the ink limit of LK higher, all else 
being equal
> it will print darker at peak and hence be 'partitioned' at a higher
> percentage of K.
> 
> 


In that case, it doesn't make that much difference (within reason). 
As long as you set enough density and overlap to cover the paper with 
no visible dots you should be fine. The only thing you might want to 
think about is how much ink is going onto the paper at a given point. 
With some inks, it is easy to get more ink than the paper can hold. 
This normally comes up with colors where you may be mixing three inks 
together to form the target. QTR may take this into it's calculations 
and compensate to prevent excess ink on paper.

Example:

You have a target (input) color of 40 C, 40 M, and 40 K (120%). Your 
3 color limit is set at 150%, because that's all the paper can hold, 
should be safe right? Now if your printer needs to use 30% LC, and 
20% C to make that portion of the color, and 40% LM, and 30% M, and 
40% K (total 160%), you are in trouble because the limiting may only 
see the 120% (the input value) and bleeding or puddles may form. The 
darker the full shade and the lower (more dense) the cross over 
point, the more light shade you need to fill in the dots. And with 
primarily only 2 inks going down at any one time, you're chances of 
going above what a paper can hold are greatly reduced. his example 
can easily happen when you linearize each channel first, and then 
have to match the light ink into the full ink's linearity.

Now I have seen one RIP that suggests you should not use more than 
about 50% of the available density from the light inks, just to try 
and avoid the above problem. I don't think that rule can be applied 
to every ink set though. And that was after a global ink reduction 
step, my printer runs best at around 60% maximum, or 40% global 
reduction, with individual limits done after that.

Steve Kale wrote:

>For those gurus who understand greyscale partitioning way better than I do I have the 
>following question.  Let's say you printed QTR's ink separation page and determined that you 
>would limit the K ink at 50% but noticed that your LK and LLK inks continued to show solid 
>increases in density well above 50%.  One could either limit each ink at 50% and calculate the 
>partitioning points based on this ink limit or decide to limit K at 50% but run LK/LLK at a 
>higher ink limit and calculate partitioning points based on this mix of limits.  Is there 
>anything to be gained by the latter?  I was thinking that perhaps one would ordinarily prefer 
>to run inks close to their "peak" densities in order to get a less "dotty" greyscale. Does this 
>make sense or does it not really matter?
>
>Cheers
>
>Steve
>
>  
>
It is flexible in the way QTR does it so you can decide what is more 
important. Roy's QTR may be more intelligent on ink amount than what I 
have done with another RIP and a more less DIY quad ink set + odd 
choices for the distribution of that inkset on the RIP's channels.

For example there's more than the Max-out point to decide where K's ink 
limit should be. Bleeding can be the limiting factor instead of the 
highest density and the ink limit should be put lower to keep contrasty 
detail possible in the shadows while sacrificing some density. With 
Photo Black the heavy gloss differential/bronzing usually lies at the 
same point of Max-out as the density drops with the loss of gloss, it is 
easier there. I also think that if the factors above allow the use of 
the Max-out patch you might gain some density by taking a few 
percentages above that patch to even out any reduction by the 
linearisation - profiling step. That's more a suspicion than a fact but 
I have done that on the Wasatch SoftRip and I didn't see strange 
deformations in the linearisation curve near the deepest black (for 
example a heavy cut in the curve just below black) .

If there's a boost with the next grey ink on the black ink you may 
consider to limit the K ink a bit more as bleeding could increase, the 
ink limit of the next grey ink will not influence the boost as it is 
just a small percentage of that ink. Why using a boost when it doesn't 
increase density and may cause bleeding ?  Not all printers print a 
smooth BO at the highest density, my 9000 with MIS Eboni doesn't, so I 
have PK running as the next grey ink and  get the black smoother that 
way. Not really a boost but good for the quality.  BTW, the MIS Photo 
Black on PhotoRag has a very short  usable grey range at 100% inklimit, 
Dmax starts early and bleeding too.

At the partioning points of the grey inks (quad sets) you may get 
bleeding problems (loss of contrasty details) faster while using all the 
steps of your lighter greys (say an ink limit up to 75%) than when you 
switch to a darker grey ink at 60%. Especially with black lines in 
highlights you may see that problem, the medium of the light grey ink 
dissolves the black ink at the edges. On the other hand the lighter grey 
ink will keep more gloss in gloss printing and a lower ink limit may 
give the visible dots of the next darker grey ink.

Toner ink is another factor. One Blue cool toner like I have does 
influence the darker warm grey patches much less than the lighter grey 
patches. The toner curve then has to be very progressive to get a smooth 
hue over the entire range. That's impossible as the bleeding gets too 
high at 70-80%. The result is that cool on this printer can't be that 
cool. The only other choice is two blue toners, one darker and more 
saturated, the other lighter and less saturated. With Sepia it goes 
better as it suits the warmer grey inks better and the saturation can be 
higher without the dots getting visible. I have mixed both for this 
printer so this isn't a general issue and a none issue on the 9600's etc 
with the UC inks.

I'm using the Wasatch SoftRip targets with the contrasty white and black 
lines in the patches to find out where bleeding is acceptable and where 
it is too high. I should adapt the QTR targets with similar lines but it 
works different there. With the typically CMYK based SoftRip I have 
ramps for Eboni + Middle Grey ink (Mm), Photo Black + Light Grey ink 
(Cc) and cool + blue toner in K + Y.  The ink limit + linearisation 
target has the ramps for the CMYK channels so two grey ones in my case + 
two toner ones. Next to that I have two build ups: of all the grey inks 
= CcMm and CcMm+K (and should have one for CcMmY but can't because that 
triggers black generation so doesn't work with that CMYK RIP). In the 
build ups there are contrasty white and black lines to check bleeding. 
In the first tests I check for bleeding in the single channels and take 
the inklimits per channel (overall number). In the second test with the 
first ink limits + partition curves I check for bleeding in the build 
ups and linearise on that target if acceptable. If not I adjust 
inklimits individually with the correction curves. The next target can 
be used for linearisation then. On top of that a Lab like curve was 
added (to my eyes).

Roy's QTR is probably more intelligent in ink amount control however it 
would be good to add the channel build ups to the target (selecting the 
channels by the user) and have the black and white lines running through 
the patches. I have not had time to check the latest incarnation of QTR 
and maybe this all has been added already.

Ernst

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Steve Kale <stevekale@b...> 
wrote:
> 

> 
> Print the calibration page at 60 then at 80.  Measure LC off the 80 print
> and compare with K on the 60 print.

Thanks Steve... :-)

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Ernst Dinkla 


Pretty much beat me to this, must have been typing when this came in or 
I might not have bothered with what I wrote. I think yours might 
explain things better.

Ernst Dinkla wrote:

>
> With the typically CMYK based SoftRip I have 
>ramps for Eboni + Middle Grey ink (Mm), Photo Black + Light Grey ink 
>(Cc) and cool + blue toner in K + Y.  
>
Should have been:

With the typically CMYK based SoftRip I have 
ramps for Eboni + Middle Grey ink (Mm), Photo Black + Light Grey ink 
(Cc) and cool + sepia toner in K + Y.  

Ernst

Steve

I'm the author of the QTR user Guide. I don't pretend to be an expert. I'm
learning as I go and simply writing down what I've learned so others can get
up to speed more quickly than I.

What I told Jenny what I thought at the time based on my experience to date
and on the comments I had received from Daniel Staver. I would be happy to
hear what benefits you gain by extracting a bit more precision from the ink
characterization step. I'm always interested in refining my approach, and
adding what I've learned to the guide.

Of course, what might provide benefit for a 4000 or 4800 may be moot for a
2200, which is what I'm using.

Tom Moore

> -----Original Message-----
> From: DigitalBlackandWhiteThePrint@yahoogroups.com
> [mailto:DigitalBlackandWhiteThePrint@yahoogroups.com] On Behalf Of Steve
> Kale
> Sent: Wednesday, July 13, 2005 12:13 PM
> To: DigitalBlackandWhiteThePrint@yahoogroups.com
> Subject: Re: [Digital BW] Re: Partitioning greys - an ink limit question
> 

....

> 
> 
> I am sorry but I don't know who Tom Moore is. Sorry Tom. At any rate, this
> is the question I am thinking about.  From a very basic level, it would
> make
> sense to ensure maximum ink coverage throughout the scale and that would
> imply running the lighter inks further up the scale (then you get your
> 'mid-tone' density through more ink rather than less).  Whether from a
> practical point of view it makes a difference I don't know.  But if people
> see benefit in more than 3 greys then this would imply that there is
> benefit
> in at least using 3 fully.
>

> What I told Jenny what I thought at the time based on my 
experience to date
> and on the comments I had received from Daniel Staver. I would be 
happy to
> hear what benefits you gain by extracting a bit more precision 
from the ink
> characterization step. I'm always interested in refining my 
approach, and
> adding what I've learned to the guide.
> 
> Of course, what might provide benefit for a 4000 or 4800 may be 
moot for a
> 2200, which is what I'm using.
> 
> Tom Moore
> 

Tom, I appreciate all your help.  Without your tutorial I don't 
think I would have even attempted to use QTR at all.  Since I am 
redoing a curve (and starting a new one) I am going to try the 
pickier method of comparing the lighter inks at 85% to the K at 65% 
and see what happens.  It may not make a bit of difference.  I just 
completed my first step wedge, will let it dry, then will try 
linearization and see how it goes.

Thanks again,
Jenny

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "byushooter" <jellerbe@j...> 
snip
> Since I am 
> redoing a curve (and starting a new one) I am going to try the 
> pickier method of comparing the lighter inks at 85% to the K at 65% 
> and see what happens.  It may not make a bit of difference.  I just 
> completed my first step wedge, will let it dry, then will try 
> linearization and see how it goes.
> 
> Thanks again,
> Jenny

Jenny, I'd appreciate it if you post what differences you find....
Ed

Hey no worries.  It is great of you to spend the time writing the tutorial.
On a first read, Ernst's post went way over my head but I intend to read it
again in the morning with a fresher brain... Thanks Ernst I am still chewing
:-)

> From: Tom Moore <r.t.moore@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Wed, 13 Jul 2005 16:44:40 -0400
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: RE: [Digital BW] Re: Partitioning greys - an ink limit question
> 
> Steve
> 
> I'm the author of the QTR user Guide. I don't pretend to be an expert. I'm
> learning as I go and simply writing down what I've learned so others can get
> up to speed more quickly than I.
> 
> What I told Jenny what I thought at the time based on my experience to date
> and on the comments I had received from Daniel Staver. I would be happy to
> hear what benefits you gain by extracting a bit more precision from the ink
> characterization step. I'm always interested in refining my approach, and
> adding what I've learned to the guide.
> 
> Of course, what might provide benefit for a 4000 or 4800 may be moot for a
> 2200, which is what I'm using.
> 
> Tom Moore
>

> snip
> > Since I am 
> > redoing a curve (and starting a new one) I am going to try the 
> > pickier method of comparing the lighter inks at 85% to the K at 
65% 
> > and see what happens.  > > Thanks again,
> > Jenny
> 
> Jenny, I'd appreciate it if you post what differences you find....
> Ed


Ed,  I am a novice at all of this so you shouldn't put a whole lot 
of stock into my observations.  My initial step wedge after trying 
the above method is that it is a bit smoother than the first one I 
did (photo rag using 65% as my ink limit for all 4 piezo inks).  In 
particular, the area around 70-80% is smoother.  I have no idea if 
it is because of the different ink limits or because I did a better 
job the second time around.  The mid-tones and highlights are very 
slightly darker so I am adjusting those with the gray curve numbers 
to see if I can get a better match to the ideal density chart.  If 
anyone else tries this I would be interested in hearing how it goes.

Jenny

From my experience it helps to select ink limits where the darkest 
patch for each ink has good separation from the next darkest patch. 
This will not be the darkest that the ink will print, but will help 
eliminate flat spots in the grey ramp. This seems to apply to the black 
ink as well and is made up by checking the 100% black ink overide for 
the the very bottom of the scale.

So I guess what I'm saying is that by not maximizing ink coverage, one 
helps the natural linearity of the grey ramp. Even with a four ink 
1160, I need a  loupe to see dots in the midtones with this approach.

-bruce


On Wednesday, Jul 13, 2005, at 10:50 US/Pacific, 
DigitalBlackandWhiteThePrint@yahoogroups.com wrote:

> Message: 15
>    Date: Wed, 13 Jul 2005 17:13:03 +0100
>    From: Steve Kale <stevekale@...>
> Subject: Re: Re: Partitioning greys - an ink limit question
>
>  From a very basic level, it would make
> sense to ensure maximum ink coverage throughout the scale and that 
> would
> imply running the lighter inks further up the scale (then you get your
> 'mid-tone' density through more ink rather than less).  Whether from a
> practical point of view it makes a difference I don't know.  But if 
> people
> see benefit in more than 3 greys then this would imply that there is 
> benefit
> in at least using 3 fully.
>