I think the major reason why QTR prints out of the box can look a little
flat is because of the combination of two key factors: (1) the relatively
weak dMax (the maximum density (or least reflectance) generated by 100%
black) of matte papers, around 1.6-7, and so the available contrast is
relatively low, and (2) the way QTR does its linearization - not that it is
necessarily wrong - one must pick a methodology - but because people don't
understand it. The first is a fact of life, the second needs to be
understood and so compensated for if deemed necessary.
If you want to see on screen the effect this weak dMax and QTR's
linearization do the following. Open a Gray Gamma 2.2 step wedge. Add a
Curves Layer (Layer->New Adjustment Layer->Curves->Enter). Grab the bottom
left anchor point and slide it up to (0,41) and grab the top right anchor
and pull it down to (255,243). Now look at how the input/outputs for the
points in the middle have changed eg 128. For the most part, the image is
lighter because the better part of the straight line has been shifted up
(128 has gone to 142). (Note that there is still nice visual separation
between the steps on the screen which is QTR's linearization goal.) When
people say an image is flat it is largely, I believe, because all the
mid-tones have been brightened.
There is no getting around the low dMax short of printing to a paper or with
an ink that leads to better blacks. So what can you do with what you are
stuck with? Ultimately it involves compromise. If you look at point 128 in
the curve above and trace across to its output you will see it, 142, is
halfway between 243 and 41 - input values are spread evenly over the
available output range. If you pull 128 back down to (128,128) your
mid-tones head back towards where they were before but there is greater
compression "in the shadows" than "in the highlights". (In essence, when
Paul is doing his curves he targets this midpoint and pulls it back a bit
leaving a little more compression in the shadows than the highlights.) Ok
now drag the 128 point off the curves chart so that it disappears - you
should now have a straight line again. Hit OK. Leave it visible.
Add another curves layer. Lock (128,128). Grab (63,63) and pull it down a
little say to (63,43). Grab (192,192) and shift it up a little say to
(192,213). This is what is meant by an s-curve - it is shaped like an s.
You are adding a little punch to the highlights and darkening the shadows a
little without shifting the mid. You can view the effect of the second
curve by itself by turning the visibility of the first off. The effect on
the image though will be the two curves combined. The selection of points
and how much to move them is really an artistic one but you have to judge
the two together which is why I said keep the first curve visible. Note
that the first curve (a representation or simulation of what the RIP is
doing) shifted the mid up point up and the second left it the same - ie
still up. So your mid grey will still be lighter than the original image
without the curves. You might choose to pull the mid darker in the second
curve to counteract the effect of the RIP. It's up to you.
If you don't use a soft-proof, adding the first curve is a good
approximation for QTR's built in transformation for printing to matte paper.
Put this curve up first, add and adjust the second so that the image looks
to your likening. Before printing, though, delete the first curve. You
don't want the effect to occur twice - it is already done by QTR and your
curve was there so you could see the effect on screen and so design your
adjustment curve.
Show quoted textHide quoted text
> From: bradmichael2003 <brad@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Thu, 16 Dec 2004 18:43:51 -0000
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: [Digital BW] Re: roark vs qtr - newbie q.
>
>
>
> I am fairly new to this, and have also at times felt that my QTR
> prints are also a little flat looking.
>
> Would you describe adding an S curve, in PS to help with this if you
> don't mind taking the time?
>
> Please pretend like I do not know anything, which at times feels like
> a pretty accurate assessment.
>
> Thanks!
>
> Brad
>
>
>
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, Steve Kale
> <stevekale@b...> wrote:
>> In fairness to QTR you will have no problem with the rear feeder if you
>> follow the Epson instructions properly - it has nothing to do with
> QTR! (I
>> have used it without problem with OS-X for two years.) The "flat
> look" can
>> be managed with an S curve in PS. Paul effectively embeds an S
> curve in his
>> calibration of his curves.
>>