Loading ...
Sorry, an error occurred while loading the content.

Re: [PanoToolsNG] Re: The 100/24 challenge

Expand Messages
  • Bjørn K Nilssen
    ... It certainly wouldn t work with my system with circular fisheye mounted on a non-dSLR, where I can vary the zoom. It is so heavy that when it is mounted on
    Message 1 of 14 , Apr 4, 2007
    • 0 Attachment
      On 3 Apr 2007 at 19:47, Yuval Levy wrote:

      > Bjørn K Nilssen wrote:
      > > Interesting concept, but I can't see how it could possibly be easier/faster to make panos
      > > in a 3D program? It would anyway need to use a very precise panohead, and a very
      > > consistent setup?
      >
      > I think his whole point, like mine, is to extract maximum value out of a
      > certain set of gear. Unlike my process which is built for fault
      > tolerance, his seems to be built for precision. From what I can see, his
      > panohead is extremely precise and very consistent.

      It certainly wouldn't work with my system with circular fisheye mounted on a non-dSLR,
      where I can vary the zoom. It is so heavy that when it is mounted on my NN3 head I have
      to use a different cropping circle on the zenith shot, because the lens apparently sags a
      little ! But, with this system I can set the zoom to any value, tilt the camera to any
      degree, shoot single or multi-row etc, and they can all be sent to PTgui and be processed
      and rendered in usually only a few minutes. And I don't even have to level the panohead..
      With my typical 4-around CF (or even 3-around) I can't see how each shot could easily be
      "defished" and mapped onto a flat plane with 120 deg fov wth any quality at the corners?
      Doing it the other way around could be very useful though :) Mapping a spherical pano to
      any (preferably mostly convex) 3D object would allow you to unwrap and extract each poly
      for making a physical Philosphere or other type of object.

      > > Besides, after downloading a 115MB mov (3D_world_us.mov) file I certainly expected more
      > > than 1:53min of small QT video !!
      >
      > he obviously is a prefectionist - i can't see any compression artifacts
      > in the above mentioned video. i am no expert on video, so i can't judge
      > wether 115mb for such a movie is too much or not, i was amazed at seeing
      > the pictures in the 3D world. IIRC you are very knowledgeable of 3D
      > software, so you can answer my question: I saw him moving the white
      > center point where the red/green/blue axis come together. Would it be
      > theoretically possible, with something like that to slightly correct
      > camera positioning?

      Why?
      Theoretically, I guess it should be possible, because it's fairly easy to slide the image
      around on each face. Moving it slightly on one face would require to adjust all the
      others faces accordingly. If you move the camera inside the panorama "sphere" off center
      you will get distorted panos, which may be like correcting camera position?

      > My impression, when looking at that, is that he is very far advanced. It
      > reminds me of Photosynth <http://phototour.cs.washington.edu/>

      Well, I don't know how advanced he is, but using 3D to assemble panos IMHO is not the
      future - not now that we have all those great pano tools with better control. I don't
      really think Photosynth is the future either, although it's an interesting concept. I
      believe more in Gogle Earth with 3D models, and either linking, likewwp, or viewing panos
      right inside GE (requires a SpaceNavigator until GE gets a pan mode). Like shown at
      http://digitalurban.blogspot.com/2006/12/digitalurban-announces-panorma-viewer.html
      and http://digitalurban.blogspot.com/2007/01/panoramas-inside-buildings-in-google.html

      --
      Bjørn K Nilssen - http://www.bknilssen.no - panoramas and 3D
    Your message has been successfully submitted and would be delivered to recipients shortly.