I've not seen any reports on the new stitcher sub-version apart
from a passing reference to the new 5.6 being out. Well, here's my
take on the 5.5.2 update that came out in May.
It's fast. Well, compared to the last version, anyway.
My typical projects are build with this:
Source: 3x12 + zenith. Nadir = nada ;)
Camera/Lens: EOS 20D, Sigma 18mm
Blend: Smart - Sharpen: None - Interpolation: Mitchell
Output size: 18000 x 9000 = ~600 MB incl. alpha channel
Machine: G5 2x2 Ghz, 3 GB RAM, 10.latest ;)
Render time:
With every advancement in blending quality, the blending steps have
consumed more and more time. From 3.5's fade-over blending that was
rather fast (1 hour or so) but sometimes hideous, until 5.5 that
took well over 8 hours to complete (for me, YMMV).
Now, with 5.5.2, it's back down to about 2 hours. THAT's what i
call progress. <cue happy dance here>
Fisheyes (7200x3600, 6-shot, Sigma 8mm) complete in about 10
minutes. Nice. Switch over to read some mail, and the stitch is
done. :)
5.5.2 is Universal Binary. Can't comment on any speed gains by
this, since i'm all PPC when it comes to Stitcher, at least for the
next weeks.
The interface has degraded a bit. I used to have the thumbnails as
a separate window. This has become a bit unsightly with the
filenames barely readable below the images, so i've reverted to the
"default" method of a line of thumbnails creeping along the bottom.
Me No Like, but with the rather good auto-stitch, it's no big deal.
Force-Stitch is gone. Instead, the use of the "manual stitch" where
one puts markers on identical points in two images is the new
"force". Had to use it only one or two times, and it worked pretty
good. Note: Instead of "force stitch", it's also possible to just
leave the image in the desired place, unstitched.
Well, that's it. Had no crashes or other problems, so, for me,
5.5.2 is my new tool of choice. 5.6 got marked "if i have money to
spare".
Reassuting note to Apple: a new MacPro is waaaay higher on my list ;)
-Markus
_______________________________________________
