1) PK RAW profile from QTR is perfectly linear, as for my eye, looking at concentric grayscale circles test charts. Tetenal satin/pearl papers.
2) Applying linearisation with use of Q13 test chart, scanner and QTR-StepWedge-Tool.zip makes the profile not only too dark as described in my original post but also non linear. Clearly visible bumps on concentric grayscale circles test.
3) I repeated the procedure of linearisation with studio flash lamps (5500 K) and digital camera (EOS 6D) instead of my scanner. RAW file, PS CC 2015 processing: results very DIFFERENT than using my scanner. Shape of the curve is like for different type paper. In theory it should give the same results since I cmpare Q13 reference with printed 21 step wedge.
4) Surprisingly, using scaner, numerical results of the output of L as computed by QTR-StepWedge-Tool are very linear after linearisation! But my eyes sey something different...
Conclusions: there is no conclusion, I am totally messed up, maybe QTR-StepWedge-Tool code has some bugs? Any ideas that would help me? Certainly, I make a mistake somewhere, but can't find it.