[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
On Thu, 5 Mar 2009, Simon Williams wrote: > Ralph Smithen wrote: >> I took it to be either that the depth of field was greater than should >> be possible (though people furthest in the crowd are going out of >> focus), perhaps indicating artificiality. > > There's a terrible join line in the middle. Two people merged into one. I think the technology's quite good - one camera on what's basically a pan/tilt base. I guess the camera is set to a good degree of zoon, then you just X/Y plot the scene to the limit of the mechanism, feed the images into a stitcher and off you go (with the occasional artifact, but hey..) I'd expect the controlling software to know the zoom level of the camera to know how far to move it for each shot, but I'd be surprised it it took more than a minute to do a huge scene like the one linked to. The site has a few other good images - one of London at night where you can zoom right into the time of the clocktower (Big Ben) For Obamas one, the image is: 59,783 X 24,658 pixels. My 5MP camera gives images of 2560 x 1950, so using that as a base, the image is 24 x 13 or 312 tiles. That would need just over 5 frames a second - might be do-able, obviously less with a bigger sensor camera... Heres a good one: http://gigapan.org/viewGigapan.php?id=3768 How many bunnies can you spot :) Gordon -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html