For anyone who was afraid to live on the bleeding edge, I just noticed that the Hugin project has released stable binaries for windows (32 and 64 bit) and Mac OS. This program is just awesome for panorama stitching. I prefer it to CS4's stitching routine any day and it's free.
stable Hugin binaries released(5 posts) (4 voices)
I've been rolling my own from SVN for the last couple of weeks and have had no issues.
For anyone who feels technically challenged, and has more money than time, I still think PTgui deserves a look. Keep in mind that at its core it uses the same math libraries as Huggin, but I feel it has some smarter defaults and results in more success with problematic (read parallax problems) images out-of-the-box. Huggin can be whipped into shape to accomplish the same goal, but again more dollars than time is what I'm talking about.
I downloaded it! i am interested to see how it competes against my CS5. Ive been holding off buying autopano pro for a while
That advantage of Hugin over photoshop is that you have some control over the stiching process, whereas photoshop gives you next to nothing. Normally I find photoshop ok, but occasionally it will fail to merge images, and doesn't leave me with many options to "help i out" so to speak. At least with Hugin, I can specify control points between objects I know should match up.
By the way, does anyone happen to know where to get the control point generators for OSX version? when I last installed Hugin, I got a note saying that these were intentionally left out for copyright reasons, but are available elsewhere (without specifying where of course).
jerl, am I wrong in believing that the main reason they've finally released official binaries is that they've got a free (as in freedom) control point algorithm now? I think they don't include Autopanosift-C in the binaries because of the potential free as in freedom problems, but they have something else that has taken its place.
I think you might be okay now with this most recent version. I had the same issue with the last version I downloaded (Dec 2010), but I'm off to try the official windows binary now. I'll update when I've checked it.
EDIT: I think I was right about the control point algorithm. I'll bet you could add autopanosift-c if you need it, but they've now got a routine called hugin cpfind that does the control point finding. It seems really fast and maybe better than autopanosift-c based on my single data point—a vertical pano that was giving me trouble when I tried it before if I'm remembering right (I know I never got a result I was satisfied with).
You must log in to post.