On 11 June 2010 02:34, Tom Sharpless <tksharpl...@gmail.com> wrote: > > OK, now I see the (multirow/stacked) CP option. > > I tried it out on an 8-shot fisheye spherical pano. It ran pretty slow > (mainly because of the poor k-d tree implementation in autopano),
Also reading and writing the .key XML files takes up a lot of the time. > could not place the zenith and nadir shots, and aligned the others > badly, but with the right arrangement. The generatekeys/autopano pair doesn't deal well with fisheye images. So this multi-row feature is currently only really suitable for narrow field-of-view photos. > Perhaps it is not fair to test a 'multirow' algorithm on an image set > with one 'row' and two 'caps'. But a practical 'grid' method ought to > handle spherical arrangements as well. It will, but the plan is to add fisheye support to Pablo's libpanomatic tool rather than continue with the distribution problems of autopano-sift-c. Do try it with a real multirow project. My experience with a 5x35 'gigapixel' project is that it just works, and although it isn't fast, it scales linearly with the number of photos. In the past you proposed a binary memory-mapped format for temporary .key files, this would improve the speed considerably. Antoine Deleforge will hopefully tackle some of these issues with his Summer of Code project. It would be a great help if you could follow the progress and chip in where necessary. -- Bruno -- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
