Digital BW, The Print

I'm not the most experienced QTR profiling person, but I've been seeing what
might make the process faster and more accurate from my perspective.  Here
are a couple of things I've found that might become part of my general QTR
profiling procedure.

 

First, I think a smooth are relatively linear carbon core is needed.  I
think a linear core makes for more even toning and a more linear final
result.  I was not satisfied that the QTR controls in the gray tabs were
sufficient.

 

So, what I've found works rather well is to use QTR's partitioning
procedure, but when the Create Curve button is pushed, use a screen grab to
copy the curves.  Then go to Photoshop and simply copy those curves'
coordinates to *.acv curves.  The PS curves can then be rather easily
refined to achieve an excellent carbon core.  These PS curves are imported
back into QTR for the individual gray inks. The resulting linear (even
before final linearization) core is then used for all the profiles for that
type of paper.  

 

This basically gives us perfectionists the advantages of Roy's partitioning
algorithm as a starting point, but with none of it's limitations.

 

The second thing I've found very useful is to set the LC and LM (toner) ink
limits by simply printing a few 50% patches with straight line curves for LM
and LC.  This is very fast.  The straight line curves are simply entered
into the Points List tab as (0, 0) and (100, 100).

 

Once the 50% patch is the tone I want, the curve is simply turned around at
about 85% to 90% so that the curve ends at (100, 0).  By getting the light
inks out of the 100% patch, the dmax is increased about 0.05 - worth the
effort.  I use *.acv curves to be sure the turn around is smooth.  I stick
some points at (25, 25), (50, 50), and (75, 75) to be sure the "curve" stays
straight for the most part and the same as the (0, 0) to (100, 100) initial
test curve.  The PS curves can be used by QTR or the points can be put into
the Points List tab to make future modifications easier.  (It's easy to lose
track of PS curves that are not accessible from the QTR profile itself, and
other users would not have the originals.)  Many will simply linearize these
basically straight line curves and be done with it.  Others (like me) will
find tweaking a bit worth the effort for some purposes.

 

Paul

www.PaulRoark.com <http://www.paulroark.com/>  

 



[Non-text portions of this message have been removed]

Paul

Could you elaborate on your method for refining the .acv curves in photoshop
and your method and criteria for evaluating them?

Tom Moore

> -----Original Message-----
> From: DigitalBlackandWhiteThePrint@yahoogroups.com
> [mailto:DigitalBlackandWhiteThePrint@yahoogroups.com] On Behalf Of Paul
> Roark
> Sent: Thursday, April 19, 2007 3:38 PM
> To: DigitalB&WPrint
> Subject: [Digital BW] QTR profiling and partitioning
> 

...

> 
> 
> So, what I've found works rather well is to use QTR's partitioning
> procedure, but when the Create Curve button is pushed, use a screen grab
> to
> copy the curves.  Then go to Photoshop and simply copy those curves'
> coordinates to *.acv curves.  The PS curves can then be rather easily
> refined to achieve an excellent carbon core.  These PS curves are imported
> back into QTR for the individual gray inks. The resulting linear (even
> before final linearization) core is then used for all the profiles for
> that
> type of paper.
> 

...

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

...

Hi Tom,

>Could you elaborate on your method for refining the .acv 
>curves in photoshop and your method and criteria for evaluating them?

Yes, I might even be able to get to it tomorrow.  (Hopefully before I
forget.)

(I'm also trying to figure out a 51-step image that works for linearization.
More steps should be better if the artifacts can be controlled.)

Paul
www.PaulRoark.com

What is meant by the term "carbon core"?

Thanks, I'm taking a deep breath, hunkering down and getting ready to 
learn this QTR profiling process.

Paul
I use a crop of the 51 step wedge that comes with the QTRip 
installation in the OneEye folder.  Is there something that you've 
found wanting with this wedge?

Jeff

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark" > 

(I'm also trying to figure out a 51-step image that works for 
linearization.
> More steps should be better if the artifacts can be controlled.)
>

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, 
"sanfo2003" <SandyCornelius@...> wrote:
>
> What is meant by the term "carbon core"?
> 
> Thanks, I'm taking a deep breath, hunkering down and getting ready 
to 
> learn this QTR profiling process.
>
I'm so glad someone else asked that!

Now "graphite core" I would have known, but I think it is very very 
different.

Ken

Hi Jeff,

>I use a crop of the 51 step wedge that comes with the QTRip
>installation in the OneEye folder.

OK, I see it's still in the package.

> Is there something that you've found wanting with this wedge?

Yes, see http://www.paulroark.com/BW-Info/51-Step-Lab-A-B.jpg 

The pattern of the 51-step file appears to be picking up artifacts.  The
artifacts may be aligned vertically on the paper -- perhaps  caused by
rollers or otherwise in the coating.  The patterns seem repeatable, so I
don't think I'm dealing with a random noise level issue, but it may have the
same impact.

I think the pattern that the 51-step file is arranged in and how it is used
may be accentuating a problem.  I'll be trying a vertical, sequential
version of the file later.  

Frankly, I'm still finding the 21-step a better compromise.  I may use a
dual vertical one for a finals.

Paul
www.PaulRoark.com 


_____________


--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark" > 

(I'm also trying to figure out a 51-step image that works 
> for linearization.
> More steps should be better if the artifacts can be controlled.)

> What is meant by the term "carbon core"?
> 

The short answer is that the B&W images I'm printing are mostly composed of
multiple carbon inks -- K, LK, and LLK.  So, achieving the best partitioning
of these core inks is a critical first step in the way I profile the B&W
inksets I'm currently using.  

To elaborate (i.e., the long version is ... ) in my view, and I think it is
fairly widely held, the best current B&W inkjet prints are those that have
mostly carbon (or equally stable, neutral pigments) as the image-forming
substance.  Carbon is warm, so we use color pigments to make a neutral image
-- or other tone like sepia.  We either mix that into blended inksets, or we
blend the carbon and color pigments on the paper via the printer, driver,
and software profile.

With the Epson K3 inkset and the variable tone inksets I've designed, one
"channel" is pure carbon -- no color pigments added.

With the older printers, with large dots, I had to do several things to
increase smoothness -- blend the colors with carbon to reduce their gamut
and limit the number of toners. 

With modern Epson printers, we can achieve the smoothness we need with
non-blended, generic color inks (LC, LM, & Y) and standard carbon inks (K,
LK, and LLK).  This is much of what the Epson "K3" approach does.

What I'm doing now is using that same basic approach, but in a generic way
and with QTR.  This, hopefully, will give us a powerful, simple,
transportable, general and generic B&W workflow from the basic entry level
hextone printer to the 3800 and beyond.  It's now working very well on my
220 (and 2200 and 7500 in different variants).

The starting place, I believe is to have the best partitioning of the core
of this image and system -- the K, LK, and LLK inks.  These will then form
the core of all the various tones.

Once we have a straight-line carbon core, the toners can be applied to that
same core as close to straight line "curves."  So little of the toners is
needed, being pure color, that the toned images are not that far from linear
and are easily linearized in the final profile.

The sliders in the driver will be able to shift the toner mix around the
same straight carbon core.  So, we're looking at a space with a carbon core,
and slider controls -- 3 profiles with midtone, highlight and shadows
sliders.  My experience so far is that a straight, uniform carbon core makes
this slider-accessible B&W space work best.

The carbon core description will be in each profile Point List and open for
all.

Paul
www.PaulRoark.com

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark" 
<paul.roark@...> wrote:
>
> Hi Tom,
> 
> >Could you elaborate on your method for refining the .acv 
> >curves in photoshop and your method and criteria for evaluating them?
> 
> Yes, I might even be able to get to it tomorrow.  (Hopefully before I
> forget.)
> 
> Paul
> www.PaulRoark.com
>

Hi Paul,

Just got around to trying this approach. Do you still remember how you 
were optimizing those curves in PS?

Thx,
Dario

Hi Dario,

 

Sorry I never got around to following up on this.

 

I think the context was working on the QTR curves.  What I did is simply do
a screen grab of the graph of the curves that the Curve Creator makes.  The
coordinates of that are easily put - one curve at a time - into PS .acv
curves, where your use the 0 - 100 format.  Then, for example, if you want
more of an overlap, you can do it in PS.  Or, if the automatic profiling
gave an un-linearized ramp that was off, you could correct that in PS.

 

While .acv curves can be imported into QTR, I find it very hard to keep
track of the curves if I want to do some more work on the profile later.
So, I simply copy the coordinates into the QTR Point List.

 

With the latest work I've been doing with QTR I'm finding that I make a
curve in PS that looks good, in general, and I just copy the points into the
Point List and give it a try.  Then I modify it as needed just by changing
the numeric values in the list.  I preview the curve as I work to be sure
it's smooth.  You do need to be a little careful of sharp turns or radical
curves in the preview, because I don't think it does any smoothing.  The
smoothing can cause unintended artifacts.  That is one reason to take a look
at the curve in PS first.  The final curves that are displayed when one hits
the Create Curve do seem to have the smoothing algorithms applied.  So, if
there is an artifact, you might pick it up there.

 

At any rate, as I become more familiar with QTR, I find I'm using PS curves
less.  The Point List with and preview are now working well enough that I
don't find much need for more. 

 

I hope this at least in part answered the question.

 

Paul

www.PaulRoark.com <http://www.paulroark.com/>  

 

 

  _____  

From: DigitalBlackandWhiteThePrint@yahoogroups.com
[mailto:DigitalBlackandWhiteThePrint@yahoogroups.com] On Behalf Of
met.graphix
Sent: Wednesday, June 06, 2007 5:54 PM
To: DigitalBlackandWhiteThePrint@yahoogroups.com
Subject: Re: [Digital BW] QTR profiling and partitioning

 

--- In DigitalBlackandWhit
<mailto:DigitalBlackandWhiteThePrint%40yahoogroups.com>
eThePrint@yahoogroups.com, "Paul Roark" 
<paul.roark@...> wrote:
>
> Hi Tom,
> 
> >Could you elaborate on your method for refining the .acv 
> >curves in photoshop and your method and criteria for evaluating them?
> 
> Yes, I might even be able to get to it tomorrow. (Hopefully before I
> forget.)
> 
> Paul
> www.PaulRoark.com
>

Hi Paul,

Just got around to trying this approach. Do you still remember how you 
were optimizing those curves in PS?

Thx,
Dario

 



[Non-text portions of this message have been removed]