Joost Nieuwenhuijse schrieb:
> When I run the optimizer, it shows the average control point distance
> decreasing to a certain value. But if I look in the optimized script and
> calculate the average from the lines
> # Control Point No 0:  0.389325
> # Control Point No 1:  2.51155
> # Control Point No 2:  1.09199
> ..etc, I get a different (higher) value.
> 
> Is the average which is shown during optimization calculated in a different
> way, or is it just scaled by some value?
> I suspect that the above distances are measured in the final panorama. Is
> that true?
> 

Yes, the printed value in the scriptfile is always the 
pixel-difference in the final panorama. 

The dialog displays the geometric mean value for the control
point distances. For ordinary control points, these are measured
in angular units, scaled to the width of the final panorama:
360 degrees corresponds to the width of the panorama.
This is new in version 2.5, and should improve optimizing
control points at large tilt angles. For t1/t2 control-points,
the pixel difference is used as in previous versions.
The scaling of the angular distance to the width of the
panorama enables mixing of different control point types in the
same script.

Therefore, the printed value may be larger than the value
displayed during optimization, especially for control
points at large tilt angles.

Regards

Helmut Dersch