Lens Correction and Panos?
After looking into what lens correction added (or didn’t add) to my portraits, I wondered if the lens correction being applied to each individual image before stitching the panorama would help. One would think that the lens correction would straighten out the individual images and then the final panorama won’t have that bent bowtie effect. So I decided to do a test with the following panorama.
After applying the lens correction and restitching I came out with the following panorama.
It’s kind of hard to tell, but it appears that the center horizontal of the panorama is a bit straighter. But the sky appears to have even more bend to it. So I decided to overlay the images to see how different they were.
There’s a pretty big difference and I think that applying the lens correction to the individual images before stitching does help the final image. What do you think? Is it a positive difference?







Not sure it’s the len correction or just the variablity of the software… how did you stitch the images? Did you assign the same control points both time?
Keith
Both panos were stitched using the same process. As far as control points, those can’t be the same since the lens correction moved those around. But I used Adobe Photoshop CS5′s photomerge with the same settings for each pass. Another interesting thing I noted, was that the photomerge of the corrected images took less time than the uncorrected. There could be other factors involved, but it was a significant difference.
I would like to this this test done on a cityscape where building distortion is more obvious. PS thanks for the podcasts.
The pano software I use, PTAssember as a GUI shell for PanoTools basically does the lens correction as part of the stitching, assuming you leave a reasonable overlap. I then use enblend to do the final merge, which handles the vignetting issues well. , which was taken with a Cannon Powershot point and shoot was done that way.