How accurate is the HFOV calculation in Hugin?

[Nathan Armer]

Because, I'm trying to calibrate the lens on my ipod touch 4g, and
that requires knowing the focal length and crop factor (or 35mm
equivalent focal length). The EXIF data says it's 3.85mm focal
length, but there's nothing about crop factor or FOV or sensor size,
and I've scoured the web via Google. I did find this:
http://www.extinguishedscholar.com/wpglob/?p=806, but I don't trust
those calculations, especially when the focal length has been rounded
to 3.9mm from the start. On one of the 360 degree panoramas I made,
Hugin calculated the HFOV at 44.09334 degrees for every image (all
taken with my itouch). Can I trust this figure and use it to
calculate crop factor and 35mm equivalent focal length?

And, please, before anyone says it's not worth it, know that that's a
huge part of the fun for me - figuring out how to squeeze the best
quality out of very limited hardware through shooting technique and
software.

[kfj]

On 20 Apr., 19:24, Nathan Armer <nathan.ar...@gmail.com> wrote:

> On one of the 360 degree panoramas I made,
> Hugin calculated the HFOV at 44.09334 degrees for every image (all
> taken with my itouch).  Can I trust this figure and use it to
> calculate crop factor and 35mm equivalent focal length?

I'd answer 'yes' to this question, provided you haven't made any grave
mistakes during the take. I suppose you want to calibrate your lens
for use with hugin, so you have a lens.ini file to use for new
projects. For this purpose, the figure you get from a 360 degree
panorama is perfectly sufficient. When doing a project based on this
calibration, you may get away with reusing the values all the time,
but often the stitch becomes better if you optimize for the values you
have in your calibration as well, as slightly modifying them may mask
slight imperfections in the take. But to have a good calibration is a
valuable starting point, and since your lens is fixed there should be
no or very little variation from take to take. (Simple devices do have
advantages, as well :)

On the other hand, calculating the FOV via the arcus tangens is a
perfectly valid method (I haven't doublechecked if the link you posted
points to correct maths). What isn't necessarily easy is to determine
the NPP of your lens, so your distance to the reference surface may
not be accurate enough to arrive at the proper figure. I'd say you're
better off using your value derived from the 360 degree panorama. And
if you do a few of these and see that after optimizing for FOV you
arrive at the same value every time, you can confidently assume to
have gotten it right.

Kay