[GRASS-user] r.walk, wrong symmetric results

[David]

Dear list,

I've tried many times to calculate movement costs with r.walk (different 
DEMs, but mostly GeoTiff format) but every time I get 
symmetric/euclidean results (radial structure of the cost surface). 
First I want pure walking information, so I set lamda = 0. The friction 
surface should now be ignored in the calculation, no? The DEMs I use do 
have valleys and ridges, so I don't understand the results of the r.walk 
calculation. Am I'm doing some dumb error?
Unfortunately it's very hard to find software which calculates movement 
cost, besides of GRASS I only know ArcGIS with the Vertical Factor in 
the path distance calculation, but the calculation does only work for 
quite small grids.
I'm grateful for any advice! thanks, david