Jeff, the Piezography K6/K7 master curves have already gone through the partitioning stage, although it is much different than how QTR partitions the gray inks. Since there is so much overlap in the piezography curves, and since Jon Cone's linearization process is from the 256 step target, there is no interpolation or spline function with the 21-step target like there is in the QTR linearization step. In both processes the actual place on the x axis where the inks start and stop stays the same, but the values at the y axis are what change to arrive at a linear profile.
>>I printed the second time and pasted the linearise values and checked. And it was linear.<<
Where are you pasting those second measurements? Back into the linearization tab? I bet if you printed and measured the 21x4 a third time you will find that it is no longer linear.
When you look at the Linearization tab in the ink curve window after pressing "show curve" I believe what you are seeing is the correction curve that is was applied to the curves. It looks like a straight line the second time you paste the linearize= values back into the profile because it it saying, "no correction needed".
I think the confusing thing is the name of the “Linearize-Data” script where you drop the measurement files. The Linearize-Data script does not do the linearizing. It only graphs the LAB values and creates a comma separated list of the measured LAB\_L values in a format the installer script can work with. What you seeing in the -out.txt graph from printing and measuring the 2nd 21x4 target is just the result of the linearization from the first target. I always recommend that you do print the target the second time as you did, and then measure it and check the graph to make sure the linearization worked, but that is the point where you should stop.
Richard Boutwell