Digital BW, The Print

I am doing a fast tour through QTR to get a feel for the process and I've hit a snag I can't get past.  I've scanned the 21 step test wedge with the value scale I've developed into Photoshop and measured the Lab values of each tone patch.  I've then opened the curve in the curve editor and  entered the values in the linearization boxes (there are now 40 boxes instead of the 21 shown in the Guide illustration) from upper left, across the top, back to the left side of the second row,etc until my last value is in the first box on the third row.  I've tried it 0-100 and 100-0 but I always get the following message: Invalid linearize curve -- not constantly increasing.  Curve Creation failed.

Any help on where I'm going wrong at this step would be appreciated.  Thanks.

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "fvanoaks" <lfbgoldwork@...> wrote:
>
>... I've scanned the 21 step test wedge with the value scale ... entered the values in the linearization boxes ... I've tried it 0-100 and 100-0 but I always get the following message: Invalid linearize curve -- not constantly increasing.  Curve Creation failed.
> 
> Any help on where I'm going wrong at this step would be appreciated.  Thanks.
>
You are not doing anything wrong per se; I had the same problem and was very frustrated because the values I entered were decreasing.

What I learned eventually is that even if your L-values are decreasing with each successively darker step in the wedge, the curve-fitting routine can fail.  A more informative error message would be helpful and I have suggested this.

The present error message reads:

Invalid linearize curve -- not constantly increasing
Curve Creation failed.

This leads the user to think the linearization values they have entered are not uniformly increasing or decreasing (depending upon if L-values or densitometry values were entered).  However, after some searching one learns that even if the entered values are monotonic, the resulting linearization may not be monotonic depending upon how abruptly the entered values change and (I gather) what Gray Curve Shadow, Highlight, and (other values?) are selected by the user.  It would help point the user in the right direction if the error message briefly stated something to this effect.

For example, the error message might read:

  Invalid linearize curve -- not constantly increasing

  Despite decreasing or increasing linearization values
entered by the user, the calculated linearization curve
may not be constantly increasing.  Check for abrupt
changes between entered linearization values or try
(lower or higher?) Curve Highlight or Shadow values.

  Curve Creation failed.

You can review the values you entered and see if one or more steps show only a minimal change and try to average this small step with the next one so that the steps are more uniform.  Another approach is to get better values entered in the first place.  In my case, this was getting the values from a better quality (wider DMax capable) scanner.  There is also a way to get values using your DSLR's built-in light meter for exposures. Perhaps that might work.

I believe also there is a minimum change that needs to occur between
adjacent entries.  I'm not sure what it is, but if the curve is too flat I
think the linearization will fail.

Paul
www.PaulRoark.com


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

I was able to get the linearization to work by spreading my 21 tone Lab readings over the entire array, spacing as well as possible given 21 is not a factor of 100.  Then I filled in the remaining boxes with values spaced evenly between those on either side of any given group. I'm assuming it worked as I didn't get an error message when I hit the "Show Curve" button, it did show the same series of curves as I got prior to filling in the array, however, but, finally, I got a really lousy value scale printed out of the 21 step wedge.  I think Tom Moore is correct that very accurate values are necessary to get a meaningful linearization.  

For now I'll stick with the non-linearized curve as I don't have a densitometer.

Can anyone recommend some newer documentation for the current version of QTR than the Tom Moore piece?  Something with a little more specificity in setting up ink limits on the setup tab and in finding the best highlight/shadow/overlap/gamma settings on the gray curve tab especially.  Since I'm not working with a variable color inkset, or one with a very limited range, I don't see me doing much with T1 and T2 for a while. I'm looking to avoid muddiness in the shadows which I think is the result of one or another of the print heads muting the detail laid down by another. 

Thanks