Matthias, I'm sorry, the cam in my testscene appears to have a "headlight". I have done the evaluation trick again this time with an ambient light in the room to replace the cam's headlight. There's very little difference in brightness now. Maybe your method is pretty cool afterall.
There are a few questions that remain for me:
- how does one choose the right width/height for a pano?
- On the prev eval. jpg example there's faceting where wall and ceiling meet
(the faceting is more obvious on the full image, but I hope you can see what
I mean). Do you have any idea why this occurs? I have evaluted the scene on
1024x600 @ 4 samples.
- I still don't understand how to go from spherical image to QTVR, but
perhaps I should just read on and keep experimenting.
Regards,
Frank Bueters
>
> Thx Frank :-)
>
> Hmmm, do you have a example scene or a render :-?
>
> Maybe the error is, that you try to evaluate to early or non
> existing things. For example if you try to evaluate the
> "occlusion-shader", you must add maybe normal
> "Surface-Properties" like color and Illumination to get a
> proper result. A second mistake is possibly that you've tried
> to evaluate, let's say Illumination on a analtic sphere and
> you've used a "default" map material and not a "spherical"
> map material, then you do not have a proper material
> coordinate system for evaluating :-?
> (same on SDS with cruel or no UV-coordinates)
>
> Matthias
>
>
> ----- Original Message -----
> From: "Frank Bueters" <[EMAIL PROTECTED]>
> To: <[email protected]>
> Sent: Wednesday, January 09, 2008 8:50 AM
> Subject: RE: Antwort: VR pano's
>
>
> > Hi Matthias,
> >
> > You're a genius. Your method produces a spherical map of
> the whole scene.
> >
> > For color and brightness there are some hurdles to take
> though. Because the
> > illumination is evaluated instead of the color objects mapped with
> > illumination maps come out relatively bright while objects
> that recieve just
> > ambient light and that are not directly lit come out black.
> > All this could be overcome by putting up light and textures
> in a scene
> > twice: once for ray traced images and once for the VR presentation.
> > Depending on the scene this may be a lot of work and result
> in two different
> > outcomes.
> >
> > Thanks to the help on this list I now know what I did not
> know before: I
> > need a program that will stitch rows of prerendered images
> together and that
> > will turn them into something that a quicktime viewer
> understands. Probably
> > PTGui.
> >
> > Thanks for your help,
> >
> > Frank Bueters
> >
> >
> > > I've simply created one big analtic sphere (2000units) and
> > > there inside a small analtic sphere (1unit)
> > >
> > > Then I've assigned the colorgrid material to the big sphere
> > > and chrome material to the small sphere.
> > >
> > > From chrome material I've deleted the
> > > specular shading
> > >
> > > Then I've evaluated the illumination
> > > from the chrome sphere (see settings on the picture except
> > > that I've used as size 1024x512)
> > >
> > > I've converted the image with photoshop cs3 to jpg...
> > >
> > > (the red sphere in the image was only a testsphere)
> > >
> > > Matthias
> > >
> > > (In short: no cam in scene, only illumination evaluated)
> >
> >
<<attachment: prev eval.jpg>>
<<attachment: prev rt.jpg>>
