linewidth ad x/y displacement in kymograph macro

[mjg]

Hi,
I am trying to evaluate kymographs of ruffling membranes. So far I was not able to set line for kymograph over 1 pixel. In fact any odd number would result in the same kymograph (1slice=1px). Rectangular selection does not work. Any suggestion how to
1. make wider ROI?
2. read x and y displacement from kymograph?
So far we were trying segmented line ROI over the membrane edge in kymograph, but pixelwise it is extremely time consuming and recommended tsp macro (read velocities from tsp) does not give you info about orientation of x displacement (ie you can not tell retraction from protrusion).
Thanx,
Martin