[hugin-ptx] Re: Release 0.4 of pvQt pano viewer.
On Mon 24-Nov-2008 at 15:21 -0800, Tom Sharpless wrote: Panini's most impressive feat along these lines is an incredibly wide angle view inside St Peter's, that nevertheless appears to have almost true perspective. I have not been able to find a good image of it on the web; the best I found is: http://commons.wikimedia.org/wiki/Image:Stpetes.JPG I've compared this painting to several photos with roughly the same point of view, both panoramic and conventional; and it blows them all away. You can't see this view with the naked eye, either. In the first place it is around 170 degrees wide! If you look carefully you should be able to notice that you are seeing the floor and ceiling from different positions, but there is no clue in the picture as to how you got from one to the other. And the vertical bits between look as they might from 100 yards out in the piazza. Nevertheless the gestalt is that this is a scene you might really be seeing. [Argg, got to run and catch a boat] There are a couple of very strange and interesting things going on with this picture: 1. The reason why there is an apparently very wide field of view is that he just removed two columns from the foreground so you can see more of the aisles, cunning, here it is with the columns reinserted: http://www.flickr.com/photos/[EMAIL PROTECTED]/3061238944/ 2. The other thing is that he has a very Naïve (or clever) approach to constructing perspective. If you look at the two arches on the right, they are exactly the same just scaled differently - This is 'wrong' and suggests that maybe the middle of the drawing was created with a camera obscura or perspective machine and the sides were extrapolated. To illustrate this, you can carry on extrapolating the sides indefinitely: http://www.flickr.com/photos/[EMAIL PROTECTED]/3060301747/ This 'wrong' perspective is a nice effect as the sides don't appear as distorted as they would in a 'normal' rectilinear view, there have been several requests for this kind of modification to hugin, but it would need some sort of interactive tool. -- Bruno --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups hugin and other free panoramic software group. To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx?hl=en -~--~~~~--~~--~--~---
[hugin-ptx] Hugin svn 3564 segfaults when loading images
After I upgraded to Ubuntu Intrepid, I recompiled hugin (and all dependencies). Hugin spits out a lot warnings about deprecated headers, but compilation finishes without errors. Now, if I start hugin with a pto inputfile or start hugin on its own and try to load images, I get a same arror every time: MainFrame::RestoreLayoutOnNextResize() Segmentation fault Any ideas what to do or what is wrong? -Heikki --~--~-~--~~~---~--~~ 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 [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx -~--~~~~--~~--~--~---
[hugin-ptx] Re: segmentation fault when stitching
cri writes: Sorry for the noob question but for debugging turned on you mean adding '--enable-debug=yes' to the command './configure'. If so it didn't work. that should do it. did you do a make clean before running configure, and then install the new executable? just making sure... --alex-- -- | I believe the moment is at hand when, by a paranoiac and active | | advance of the mind, it will be possible (simultaneously with | | automatism and other passive states) to systematize confusion | | and thus to help to discredit completely the world of reality. | --~--~-~--~~~---~--~~ 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 [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx -~--~~~~--~~--~--~---
[hugin-ptx] Re: Release 0.4 of pvQt pano viewer.
Quoting Tom Sharpless [EMAIL PROTECTED]: Hi Mick Here is a very interesting paper on making strip panoramas out of ordinary photos taken from multiple viewpoints, written by one of the leaders in commercial image processing technology: http://grail.cs.washington.edu/projects/multipano/agarwala_sig06.pdf. It was inspired by the work of artist Michael Koller, whose website you should study. http://www.seamlesscity.com/ I've seen this before. I tried to make one of an Irish street of shops but there were cars everywhere and I tried to do it from the pavement in front of the shops and I didn't take enough probably it was impossible to stitch. that was pointed out as an example at the time. The slit-scan method is a far easier, hence more popular method. Here are some of my favorite examples: http://www7a.biglobe.ne.jp/~slitcamera/egindex.htm (whole trains self- scanned) http://www.danubepanorama.net/en/Main/Showcase (banks of the Danube scanned by boat) http://www.cs.iupui.edu/~jzheng/RP/Changan/ (a long city street scanned by bus) http://www.flong.com/texts/lists/slit_scan/ (a wonderful collection of slit-scan art) That last one is brilliant! I need to do some of that. I've felt a bit flat about photography recently that has cheered me up ! Friends used slit scan in animation. pan a slit across a piece of art work and move the camera in so that the image is distorted. do this progressively over frames and the artwork will spin, kind of thing. was a lot easier with the advent of motion controlled rostrums but then computers got smart enough to do it themselves. As you say, and I never thought about it before, the perspective is right from the middle of each face of the original photograph and then the equirectangular is a blend of the flat images ignoring the perspective. So I suppose more photos is better for that but I never saw it as been an issue. this is a suitable examplehttp://www.panagito.com/MISC/quorr-abbey2.html somewhere in there then the perspective should be wrong around the seams ( I think it is 6 photos around ) but I cannot see it. You can't see it because it isn't wrong. The whole point of pano stitching software is to reshape the separate pictures so they fit together perfectly on a sphere, then project the spherical composite image back into a flat one. So you won't see any sign of the perspective of the taking lens in a well stitched panorama. Nor will you see any sign of the cube edges in a cubic panorama that is being displayed by a properly designed program. The mapping from cube faces to screen is arranged to make it look as if you were viewing that spherical image from the inside, with your eye right at the center. I still haven't got this. ( I've never had to think about it because the software Just Works ) The camera images that exist have perspective so I seem to think something should go wrong at the seams. I think I need to know how the stitching works. are you saying that the rectilinear images are warped in spherical arithmetic and then distorted to fit together and then projected back to the flat equirectangular. Is the blending done on the equirectangular images ? I am not sure if what you imagine is possible just with photography I think it requires cheating one of the 3 perspective points. Something like that is one of the main ingredients in the baroque wide angle drawing technique. But there are usually more than 3 vanishing points -- and in some areas, just a subtle continuous shift of perspective. Perhaps the old masters just didn't know what they were doing :-)http://www.google.co.uk/search?hl=enq=hockney+camera+obscurameta= Hockney is probably right that some old masters used images projected by lenses to help them get shapes right -- as do many modern masters. But that couldn't possibly have helped Panini -- the lenses of his day had terribly narrow fields, and even a Canon 12mm would not have showed St Peters the way he drew it. It is a work of the imagination -- his and the viewer's. the camera lucida uses a prism apperently tho' I never saw one. And the camera obscura could be a giant pin hole camera. no need for lenses. Regards, Tom cheers midk -- This mail sent through http://www.ukonline.net --~--~-~--~~~---~--~~ 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 [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx -~--~~~~--~~--~--~---
[hugin-ptx] Re: segmentation fault when stitching
You did right by making sure! that was exactly what was needed. ;-) Thanks for the help Cristian --~--~-~--~~~---~--~~ 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 [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx -~--~~~~--~~--~--~---
[hugin-ptx] Re: Release 0.4 of pvQt pano viewer.
Hi Mick Koller says he does all the compositing by hand in Photoshop, and sometimes needs to combine 100s of shots for one city block. Not easy work! On Nov 26, 12:06 pm, [EMAIL PROTECTED] wrote: Quoting Tom Sharpless [EMAIL PROTECTED]: That last one is brilliant! I need to do some of that. I've felt a bit flat about photography recently that has cheered me up ! Friends used slit scan in animation. pan a slit across a piece of art work and move the camera in so that the image is distorted. do this progressively over frames and the artwork will spin, kind of thing. was a lot easier with the advent of motion controlled rostrums but then computers got smart enough to do it themselves. I did a few experiments with a turntable and a slit camera a few years back, that you might find amusing: http://home.comcast.net/~scancams/images-top.htm (the 3 links near the bottom). I made several such cameras from flatbed document scanners -- difficult, slow, but very high-res. A high def video camera is easier and faster -- you extract the slit scan with software. Cheers, Tom --~--~-~--~~~---~--~~ 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 [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx -~--~~~~--~~--~--~---
[hugin-ptx] Re: Release 0.4 of pvQt pano viewer.
Hi Bruno On Nov 26, 9:24 am, Bruno Postle [EMAIL PROTECTED] wrote: On Mon 24-Nov-2008 at 15:21 -0800, Tom Sharpless wrote: Panini's most impressive feat along these lines is an incredibly wide angle view inside St Peter's, that nevertheless appears to have almost true perspective. I have not been able to find a good image of it on the web; the best I found is:http://commons.wikimedia.org/wiki/Image:Stpetes.JPG I've compared this painting to several photos with roughly the same point of view, both panoramic and conventional; and it blows them all away. You can't see this view with the naked eye, either. In the first place it is around 170 degrees wide! If you look carefully you should be able to notice that you are seeing the floor and ceiling from different positions, but there is no clue in the picture as to how you got from one to the other. And the vertical bits between look as they might from 100 yards out in the piazza. Nevertheless the gestalt is that this is a scene you might really be seeing. [Argg, got to run and catch a boat] There are a couple of very strange and interesting things going on with this picture: 1. The reason why there is an apparently very wide field of view is that he just removed two columns from the foreground so you can see more of the aisles, cunning, here it is with the columns reinserted: http://www.flickr.com/photos/[EMAIL PROTECTED]/3061238944/ Aaahh! 2. The other thing is that he has a very Naïve (or clever) approach to constructing perspective. If you look at the two arches on the right, they are exactly the same just scaled differently - This is 'wrong' and suggests that maybe the middle of the drawing was created with a camera obscura or perspective machine and the sides were extrapolated. To illustrate this, you can carry on extrapolating the sides indefinitely: http://www.flickr.com/photos/[EMAIL PROTECTED]/3060301747/ Very interesting. So this is some kind of mechanical drawing projection, then, and not a rectilinear one? Just how do they differ? This 'wrong' perspective is a nice effect as the sides don't appear as distorted as they would in a 'normal' rectilinear view, there have been several requests for this kind of modification to hugin, but it would need some sort of interactive tool. What sort of tool? Let's at least design it, maybe it can be implemented. Regards, Tom --~--~-~--~~~---~--~~ 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 [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/hugin-ptx -~--~~~~--~~--~--~---
