Digital BW, The Print

What I meant was that you should do the curves as prescribed by Roy in his
notes and linearize them also according to Roy's instructions (unfortunately
this ideally requires a densimeter but I think some people have used their
scanners).  This way you can also use curves compiled and linearized by
someone else (you only need to look to see what dMin and dMax are in the ink
descriptor file).  If you don't like the linear change in L from dMin to
dMax and the density shift this implies, it is easy to calculate the curve
required in PS which, together with the "(density) curve" in QTR, will get
you the net curve you want.  I think this would be a lot easier than messing
with the inks themselves.  I just did this for my EEM and HPR curves and it
works well.

As an example from my HPR curve, if you know QTR will print value 128 with a
density normally associated with a pixel value of 144 and you want it to
remain 128 (L=50 18% reflectance grey) then you need a curve that maps 128
to 107 in the file before it goes to QTR where 107 gets printed as 128.
Calculating the adjustment curve values is just a process of interpolation.
Probably more accurate than relying on your eye which will be heavily swayed
by the surrounding shades of grey.


> From: bruce greene <bagreene@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Mon, 13 Dec 2004 21:07:56 -0800
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: [Digital BW] Re: LAB Step Wedge -- a grayscape Lab space
> 
> 
> 
> On Monday, Dec 13, 2004, at 10:41 US/Pacific,
> DigitalBlackandWhiteThePrint@yahoogroups.com wrote:
> 
>> Message: 5
>>    Date: Mon, 13 Dec 2004 16:46:09 -0000
>>    From: "Steve Kale" <stevekale@...>
>> Subject: Re: LAB Step Wedge -- a grayscape Lab space
>> 
>> 
>> Bruce
>> 
>> Rather than messing with the QTR curves you are better off using an
>> Excel spreadsheet to
>> compute a PS curve that, together with the curve in the RIP, gives you
>> the "net curve" you
>> want.  I gave you the way QTR will remap input to output.  If you
>> layout a sheet with the
>> same calcs but add a column for your target curve (graphing this is
>> useful so you can see
>> its shape as you alter the numbers) you can then solve for the curve
>> which will remap the
>> file values so that when the adjusted values are then remapped by QTR
>> you will get the
>> curve you want.  All without a densimeter.
>> 
>> Steve
> 
> 
> Thanks Steve, but unfortunately this is now going over my head a bit.
> My curves are made by eye using QTR without linearization so they are
> done, I think, using ink percentages. Of course dot gain makes the
> steps uneven over part of the output, so I don't use those ink
> percentages in my curves by limiting the maximum ink out put in each
> ink to where there are clear and even (visually) steps from darkest to
> white. My "raw" or 1st trial profiles come out with near gamma 1.8
> (differs on different papers). I then create a temporary photoshop
> levels adjustment layer to match my test image and step wedge (in gamma
> 2.2 working space, viewed on calibrated monitor) to the first trial
> curve print. I then create another temp levels adjustment layer to
> bring the monitor image back to the original look using the levels
> gamma adjustment ( I use the eye dropper to make sure I'm accurate to
> the original file). I then type this gamma adjustment into the ink
> descriptor file and create a new curve. Turning off the temporary
> levels layers, I compare the new curve printout to the original file.
> The match will be quite good. So I've now created a QTR curve for
> gamma2.2 greyscale working space.
> 
> To add a little more punch (if I desire) to the QTR curve I create a
> curve adjustment layer temporary layer to add a little punch to my
> original. I save this curve and open and apply it to QTR supplied
> "grayscale-16.psd" file. This is a miniature (pixel wise) grey wedge in
> 16bit gamma 2.2 space. I save this file with a new name with the curve
> applied and create a new QTR curve by adding the name of this new
> greyscale file in the curve descriptor line in the ink descriptor text.
> (yea this is going over someones' head!) When the new QTR curve is
> made, my photoshop "s" curve made by eye is now built in to the new QTR
> curve on top of the accurate gamma 2.2 QTR curve. No need for Excel.
> Photoshop provides the graph. It's a mystery how exactly this works as
> the instructions with QTR have "a style all their own" and it took me a
> while to figure out that I needed to actually apply the curve to the
> greywedge and save it under a new name. I guess the QTR curve making
> script examines the contents of this greyscale, notes the changes in
> tonal values and applies these changes to a new QTR curve. It's kind of
> neat actually and very cleaver of Roy I think. How the QTR curve making
> software finds this file is a mystery to me as the instructions don't
> say where (which folder) to save the file in. I saved it in the folder
> with the other curve descriptor files I think.  ***I am using an older
> version of the QTR. If the curve making aspects have changed, then no
> one will understand what I'm saying here!
> 
> I would encourage the curious who use QTR on the Mac to try making
> their own curves for quadtone ink sets. It is very educational about
> the way the different inks are laid down and the way a good transition
> works from one density ink to the next. What's really amazing is how
> well the curve making software (ink partitioning) works after one has
> identified the maximum amount of each color ink to lay down. The
> transition is so smooth, with no posterization. If you print
> photographs of people, this is really important. When I used the old
> piezo plug-in, it looked fine on landscapes, but faces always showed
> the poor partitioning of the old plug-in from cone/R9. What I mean by
> this is that if I print a grey gradient from white to black, under an
> 8x lupe, I cannot see when one inks stops and the next lighter ink
> starts. Very cool.
> 
> enough blabbing...
> -bruce