QTR-Quadtone RIP

On 29/03/10 5:49 PM, "Michael King" <drmrking@...> wrote:

> Sure I'll do some tests with the bullseye.
> I am checking the smoothness of the linearity and I didn't see any flat
> spots.
> More on this after some testing. But also its a tough image to print.
> As I said I'll upload the section that tends to posterize for people to test
> if they want.

I think I know what Michael is trying to describe. When I build 7 ink
profiles for my Cone inks and I print the bulls eye on Keith Cooper's test
print (Northlight Images) I get what appears to be "banding". The concentric
circles do not progress smoothly, instead there are light and dark bands.
I've had a number of custom profiles made by Jon's team and they conversely,
are smooth. I know that Jon's profiling tool gradually overlaps the inks. I
also know that it defines each of the 256 points for all seven inks. He
makes a "master curve" for each printer supported, which I understand is a
lot of work. Simply put, the tool is very sophisticated.

On 29/03/10 5:36 PM, "Roy Harrington" <roy@...> wrote:

> I think Shilesh has a valid point.  With the standard QTR transitions when
> you use more inks the slope of the transitions are steeper since there are
> more up and downs.  With just 3 grays the transitions are naturally wider.
> Graph Piezo K7 curves and QTR UCk3 curves and note that the slopes a pretty
> comparable.   What all this means is that you need to be more accurate
> in the relative density measurements with K7 inks and QTR curves.
> (I take it you are creating your own curves with K7 inks using QTR curves).

Perhaps its a combination of ink partitioning and getting the density of
each ink set at the right level. I've wondered lately if I've been setting
the density of each ink too high, which would lead to steep transitions.

> I must admit I haven't made many K7 curves since Cone provides them,
> but like Shilesh says with K3 inks I've never had issue with smoothness.
> What I think you are calling posterization is actually flat spots in a what
> should be a smooth gradient.

That makes a lot of sense. My profiles are nice and linear, and there is
nothing to indicate a significant problem, yet I still get this "banding". I
assume there are problems in between the 5% steps that do not manifest when
the 21 step wedge is printed.

> In the new release I've included a bullseye.tif pattern in the Eye-One folder.
> Print that out with your setup.  This shows anomalies very easily in the top
> half.  Then the bottom half makes it easy to see where the issue is.
> With a graph of the curves you can see which inks are in transition at that
> point.  If its a flat spot the lighter ink is marked lighter that it really
> is.

Sounds like a great tool. It will be interesting to see what it reveals.

Terry.