I've observed a problem with the GL preview that makes it very hard to use for many of my panoramas. Particularly when I use a wide angle lens, I find that after I do enough optimization steps the preview gets badly garbled such that it's essentially unusable. If I use a normal or short telephoto lens, this usually doesn't happen.
The sequence of steps I usually use is: 1) Create control points 2) Perform an initial optimization step, then do position/view/barrel distortion, then optimize everything except translation (lens distortion is significant with my 8-16). 3) Use the GL viewer to center and straighten the pano. 4) Calculate optimal size and FoV in the stitch panel (I hate that 3000 pixel size that all panos start with; it's one extra step I have to do). 5) Crop out foreground and sky CP's (Celeste isn't perfect, and I usually don't use a tripod much less a pano head) and re-optimize. 6) Remove all control points except between adjacent images (I shoot panos with a lot of overlap, and I've found that control points between non-adjacent images lead to larger errors in the output). 7) Crop out any foreground points that made it through the original crop, and other control points with a lot of error (carefully -- I want to make sure that the points with large error are actual bad points; if there are a lot of foreground points, sometimes the good points are the ones with large residual error). I run multiple passes, gradually bringing down the acceptable error. At this stage I usually try to get it down to about 3 units. 8) Fine tune all points, remove all points with correlation coefficient less than 0.8, reoptimize, remove the points that the fine tune botched. 9) Continue with iterative optimization until I get worst case error under 1 unit. 10) Do photometric optimization. Anyway, at some point during all of this the preview gets garbled. I found a particular case with a particular pano where the GL preview is good, but the next optimization step results in a garbled preview. The original images in question are all in http://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/ . The "good" preview is http://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-tmdKB7J; the next optimization step results in http://rlk.smugmug.com/Photography/Hugin-demo/n-NbPQS/i-vSbVSWP . This is independent of projection (at least equirectangular and Mercator). The .pto file is attached. In case it matters, I'm using the current repository (although this problem has been around for at least a few years, I think) with a Radeon HD5870 using the radeon driver with kernel 3.11.10 and the following Mesa bits: DirectFB-Mesa-1.6.3-4.1.3.x86_64 Mesa-10.2.4-381.1.x86_64 Mesa-32bit-10.2.4-381.1.x86_64 Mesa-demo-8.2.0-37.2.x86_64 Mesa-demo-x-8.2.0-37.2.x86_64 Mesa-devel-10.2.4-381.1.x86_64 Mesa-libEGL-devel-10.2.4-381.1.x86_64 Mesa-libEGL1-10.2.4-381.1.x86_64 Mesa-libEGL1-32bit-10.2.4-381.1.x86_64 Mesa-libGL-devel-10.2.4-381.1.x86_64 Mesa-libGL1-10.2.4-381.1.x86_64 Mesa-libGL1-32bit-10.2.4-381.1.x86_64 Mesa-libGLESv1_CM-devel-10.2.4-381.1.x86_64 Mesa-libGLESv1_CM1-10.2.4-381.1.x86_64 Mesa-libGLESv2-2-10.2.4-381.1.x86_64 Mesa-libGLESv2-devel-10.2.4-381.1.x86_64 Mesa-libglapi-devel-10.2.4-381.1.x86_64 Mesa-libglapi0-10.2.4-381.1.x86_64 Mesa-libglapi0-32bit-10.2.4-381.1.x86_64 libOSMesa-devel-10.2.4-381.1.x86_64 libOSMesa9-10.2.4-381.1.x86_64 libOSMesa9-32bit-10.2.4-381.1.x86_64 -- Robert Krawitz <r...@alum.mit.edu> MIT VI-3 1987 - Congrats MIT Engineers 6 straight men's hoops tourney Tall Clubs International -- http://www.tall.org/ or 1-888-IM-TALL-2 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 -- A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ --- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group. To unsubscribe from this group and stop receiving emails from it, send an email to hugin-ptx+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/201408050153.s751rZq9027345%40dsl092-065-009.bos1.dsl.speakeasy.net. For more options, visit https://groups.google.com/d/optout.
almanor-16-bad.pto
Description: Binary data
