[ptx] Hugin: strange behavior of focal length <-> FOV arithmetic

[Stanislav Brabec]

Hallo hugin developers.

I have found strange behavior of focal length <-> FOV arithmetic with my
Minolta DiMAGE A1. It looks like that for different crop factors with
the same equivalent focal length hugin gives different results. Because
in default setup Hugin believes to these values, it creates ugly seams
in panoramas.

How to repeat:
1) Take any image from
http://www.penguin.cz/~utx/Minolta_A1_ukazky/hugin/

2) You will see:
v=65.47045
focal length=7.20313
crop factor=3.8872
It gives 28.00mm ekv.

3) Save as temp.pto

4) Now set
focal length=28mm
crop factor=1
You will get
v=63.43997

With the same equivalent focal length, v is different by 2 degrees!

5) Load temp.pto

4) You will see first seen FOV and:
focal length=26.92148
crop factor=1
(it gives 26.9mm ekv.)

You will get 4% different equivalent focal length, than product of focal
length and crop factor.

5) After my experiments with upper mentioned panorama, it looks like
real FOV is 62.8411, i. e. relatively near to value returned by EXIF, if
I made computations externally and entered crop factor 1, and very far
from value Hugin computes.


________________________________________________________________________

Stanislav Brabec
http://www.penguin.cz/~utx