QTR-Quadtone RIP

I am trying to create a neutral curve for Epson Enhanced Matte paper. I am using Lyson Cave Paint pigmented Inks on an Epson 7600 printer. I started with the enhanced matte warm curve for the 7600 and followed all of the steps and did also a little experimentation. In other words, I have been all over the map with this.

Here is exactly where I am stuck. The user guide by Tom Moore says something I am not, so far, able to get. Here is a quote from the user guide:

"Once you print a step wedge, the next task is to visually assess the overall appearance of the print. Can you see all 21 distinct densities? Do they appear more-or-less evenly space. ..."

Back to me. I can see the distinct densities in the highlights and mid tones, but when it gets to the shadows, they start bunching up and are too close to each other in density. Nothing like the separation of steps in the highlights and mid-tones. In other words, the steps do not appear more-or-less evenly spaced (in the shadows.)

That's all I got, any suggestions???

Thanks...

Hi Steve,

> I am trying to create a neutral curve for Epson Enhanced
> Matte paper. I am using Lyson Cave Paint pigmented Inks 
> on an Epson 7600 printer. 

> I started with the enhanced matte warm curve for the 7600 ...

Do the Cave paint inks generally print similarly to the UC inks?  If they are close, starting with a UC profile probably will work.  You could also just print the necessary 21-step test strips to do a fresh partitioning.  You might end up with a closer fit to the ink.  But you may not want to get into this yet.

 
> ... Here is a quote from the user guide:
> 
> "Once you print a step wedge, the next task is to visually assess the overall appearance of the print. Can you see all 21 distinct densities? Do they appear more-or-less evenly space. ..."

> Back to me. I can see the distinct densities in the highlights and mid tones, but when it gets to the shadows, they start bunching up and are too close to each other in density. ...

That is one of the reasons to start from scratch.  The curve you're starting with is probably not very well partitioned for the ink you're using.

But, it might still be fine.  

I actually prefer compressed shadow tones on pre-linearization curves, as long as there is some minimum separation so that QTR doesn't refuse to partition the input.  Many of the errors I see in final curves are due to the software not having enough deep shadow information.  If the information is not there QTR will interpolate.  In the deep shadows where the curves are in steep cross-overs the interpolation may miss the mark.  Actual information as to what is going on down there really helps the system.

I prefer to read the 21-step curves with a spectro and then graph the results.  This is a far more accurate what to "see" what is happening than just eye-balling it.  I use Excel to do the graphing.  There are other programs, and OpenOffice.org spreadsheets can do the job also, though more steps are needed.

At any rate, for eye-balling the deep shadows, if you can't see into them very well even with a good light, put a strong light behind the print.  Reflective light passes through the ink twice -- transmissive only once.  So, the effective density is cut in half and it's much easier to see what is going on in the deep shadows.

Good luck.

Paul
www.PaulRoark.com

Hi Paul,

--- In QuadtoneRIP@yahoogroups.com, "pr_roark" <paul.roark@...> wrote:
>
> Hi Steve,
> 

> You could also just print the necessary 21-step test strips to do a fresh partitioning.  You might end up with a closer fit to the ink.  But you may not want to get into this yet....
 
> That is one of the reasons to start from scratch.  The curve you're starting with is probably not very well partitioned for the ink you're using....

Do I start from a completely zeroed out *.qidf file without any information entered into the fields?

> I actually prefer compressed shadow tones on pre-linearization curves, as long as there is some minimum separation so that QTR doesn't refuse to partition the input. 

The minimum separation is a problem, but I only have an Epson Scanner to scan the image then convert to Lab mode and measure the steps with the "Info" window.

> Many of the errors I see in final curves are due to the software not having enough deep shadow information.  If the information is not there QTR will interpolate. In the deep shadows where the curves are in steep cross-overs the interpolation may miss the mark.

The separation in the deep shadows are separated by steps of just one or two for about the first 5 or 6 steps from the deepest black.

So far as the linearized curve, it looks like a shallow bowl tipped up on its side. The straight line curve or at least close to a straight line, doesn't exist.
 
> I prefer to read the 21-step curves with a spectro and then graph the results. 

As I said above I only an Epson scanner and not a spectro.

Steve Law

Steve,
 
 
> > You could also just print the necessary 21-step test strips to do a fresh partitioning.  You might end up with a closer fit to the ink.  But you may not want to get into this yet....
 
 
> Do I start from a completely zeroed out *.qidf file without any information entered into the fields?


Usually you start with a Calibration Mode print that prints 21-step test strips of all the inks.  (With a 7600 and if you just making the LK-K partition, you can speed this up by making individual test strips, but that gets into how to make them.  So I'd stay at the moment with the normal procedure.)  Have the Calibration Mode print set to print at 100% limit (the slider near the bottom of the page.)

(By the way, I'm PC, so I'm not sure what a Mac interface is like.)

Use the 100% calibration curve to set you ink limits in a new, *.qidf file.  Be sure to zero out the linearization tab also.

Print another Calibration mode at the K ink limit.  The point is to get a 21-step print of the black ink printed at its ink limit.  This is used to figure the relative densities of the midtone inks.  The boxes you'll need to fill in in the Curve Creator are the ink limits and relative densities of the black and LK inks to allow the system to partition these.

(Or 21-step with a profile that is simply K 0-100 [or ink limit].  I often make straight line curves for specific purposes just to save the time of printing an entire Calibration mode print.  To do this, hid Load curve (in drop down menu just below the ink name in the Curve Creator), press Curve button, go to point list, enter (0,0) and (100, 100) [or ink limit] -- be sure it's a straight line curve.  I call name these curves, e.g., "K-0-100.")



> The minimum separation is a problem, but I only have an Epson Scanner to scan the image then convert to Lab mode and measure the steps with the "Info" window.

Scanners work fine for measuring relative densities of grayscale inks.  I don't trust them for color or absolute readings, but you don't need these at this point.  That said, I think for B&W work the ColorVision spectro is a very good tool to have and a good value.

Hope this helps.

Paul
www.PaulRoark.com