The reason why the nadir is white is because it has a different EV
value to the rest of the images, presumably because exposure
correction went wrong. Exposure correction will go badly wrong if
the images are not well aligned in the first place.
Looking at the project, the nadir also has wildly improbable values
for lens distortion which seems to be because it has been given a
different lens number and had all parameters optimised - Did hugin
give it a different number or were you trying to do something with
this shot?
--
Bruno
Ah yes, by rotating the nadir image it gets a different 'lens
number', this means that lens parameters can't be linked with the
rest of the photos and end-up being optimised separately. In this
case there was not enough information to successfully optimise
'both' lenses.
>About the lens parameters, at one point after the optimization I got
>very strange warped image, then I checked the cropping and discovered
>nadir cropping being out of the image.
>So, I tried to fix the crop by seeting the crop values manually, but I
>never succeeded to fix it as I wanted. Could it have any impact on the
>lens params?
No, in Hugin the crop is entirely a mask that is applied to the
input photo, i.e. it just controls which parts of the image are
treated as transparent and doesn't relate to field of view or
distortion parameters.
--
Bruno
Sorry, maybe someone else can help as I've never used smartblend in
Hugin?
>Another thing is the auto generation of CPs. I've never succeeded to
>get hugin do that for me, I either select all point manually or run
>autopano.sift separately, the use its output in hugin, and add some
>more point in upper and lower parts of the images and remove the CPs
>from moving objects.
Assuming that you are using Windows, I understand that the current
default of autopano-sift-C more-or-less works. What version of
Hugin are you using and where did you get it?
--
Bruno
It should be ok, I would try resetting the 'Autopano' preferences to
the default.
>(1) The strange thing (number 1) is that I get different max size
>(optimal size) almost every time (I mean for different panos but
>with the same camera/lens), not a huge difference but something
>between say 7600x3800 to 8200x4100.
I've noticed this too, but since the number is always about right, I
have never investigated the reason.
--
Bruno
>(1) The strange thing (number 1) is that I get different max sizeI've noticed this too, but since the number is always about right, I
>(optimal size) almost every time (I mean for different panos but
>with the same camera/lens), not a huge difference but something
>between say 7600x3800 to 8200x4100.
have never investigated the reason.
>Sorry to sound dumb, but does this not depend on the overlap between the
>pictures. So I would be surprised if you get the same optimal size for any
>of two different panoramas, since the overlap is almost all the time a
>little bit different.
Only for partial panoramas, for a 360° panorama the 'optimal size'
should be entirely a function of the input photo parameters. So if
you load your lens settings then you should always get the same size
panorama.
--
Bruno