Reference original image, taken with an 8 mm lens on a 1.6x crop body
(Sigma 8-16 on Canon 7DmkII) [1]
I should note that initially everything is fine -- the mapped photos
look reasonable. However, at some point during the optimization
process, and I have not managed to track down when, the below behavior
manifests itself if I'm using a very wide angle lens. I tried to
reproduce it from scratch, but wasn't successful -- I didn't find the
exact point at which things went off the rails.
It appears that Hugin is cutting off the corners of the frame [2].
I'm presuming that that's to avoid severe distortion, which is
nevertheless still visible near said corners. When Nona remaps the
images, that corner distortion is still present. Enblend seems to be
able to avoid using those areas [3], but multiblend (which is
otherwise preferable for the most part) doesn't [4].
I also note -- and I've mentioned this previously, in 2014 [5] -- that
the OpenGL preview shows some really weird errors [6], which the slow
preview doesn't [2]. However, in 2014 the corners were not cut off [7].
[1]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-24vkc6v
[2]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-Tz5fLgw
[3]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-zbWZbTC
[4]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-tzdPc87
[5]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-vSbVSWP
[6]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-SG9bZrM
[7]
https://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-tmdKB7J
--
Robert Krawitz <
r...@alum.mit.edu>
*** MIT Engineers A Proud Tradition
http://mitathletics.com ***
Member of the League for Programming Freedom --
http://ProgFree.org
Project lead for Gutenprint --
http://gimp-print.sourceforge.net
"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton