Greetings all ... I was wondering to what extent PovRay could assist with the
simulation of projection designs for immersive environments I work on. By this
I generally mean domes with one or more projectors, the delivery of digital
content requires image warping .... all of which I am familiar with in a
mathematical sense but I need a "nice" way of presenting the concepts to
"clients". As an example, imagine a dome with 5 projectors creating an
overlapping video image around the rim. See
     http://local.wasp.uwa.edu.au/~pbourke/tmp/sim/
I can convey how the images appear (see the model.zip file), but it breaks down
if I want to look at things from the outside of the dome. It also fails if I
introduce a mirror in front of the projectors (a fairly common thing in the
projection community). ps: in case it isn't obvious, the frustums around the
light source in the model act as the data projector beams, the light source is
at the apex of the frustum.

So my question is: Is there a way to use photons (albeit probably expensive) to
form the projection of images onto the dome?

If you think it is but it will take you time to demonstrate it then please feel
free to quote for the work required. Probably better to do so offline, email:
pau### [at] gmailcom

Post a reply to this message

Paul Bourke nous illumina en ce 2008-06-29 08:55 -->
> Greetings all ... I was wondering to what extent PovRay could assist with the
> simulation of projection designs for immersive environments I work on. By this
> I generally mean domes with one or more projectors, the delivery of digital
> content requires image warping .... all of which I am familiar with in a
> mathematical sense but I need a "nice" way of presenting the concepts to
> "clients". As an example, imagine a dome with 5 projectors creating an
> overlapping video image around the rim. See
>      http://local.wasp.uwa.edu.au/~pbourke/tmp/sim/
> I can convey how the images appear (see the model.zip file), but it breaks down
> if I want to look at things from the outside of the dome. It also fails if I
> introduce a mirror in front of the projectors (a fairly common thing in the
> projection community). ps: in case it isn't obvious, the frustums around the
> light source in the model act as the data projector beams, the light source is
> at the apex of the frustum.
> 
> So my question is: Is there a way to use photons (albeit probably expensive) to
> form the projection of images onto the dome?
> 
> If you think it is but it will take you time to demonstrate it then please feel
> free to quote for the work required. Probably better to do so offline, email:
> pau### [at] gmailcom
> 
> 
If you only want the projection, then a "slide" with a filtering image_pattern 
will do the trick. You can use a single faced object such as a polygon for that.
If you want to have reflections of the light beam, then you'll need to use 
photons. To help reduce the photons phase, you can set "photons pass_through" on 
the slides themselve and set the mirror(s) as photons target.

-- 
Alain
-------------------------------------------------
You know you've been raytracing too long when you look at a matrix transform and 
know instantly what it does.
John VanSickle

Post a reply to this message

>> I can convey how the images appear (see the model.zip file), but it 
>> breaks down
>> if I want to look at things from the outside of the dome. It also fails 
>> if I
>> introduce a mirror in front of the projectors (a fairly common thing in 
>> the
>> projection community). ps: in case it isn't obvious, the frustums around 
>> the
>> light source in the model act as the data projector beams, the light 
>> source is
>> at the apex of the frustum.
>>
>> So my question is: Is there a way to use photons (albeit probably 
>> expensive) to
>> form the projection of images onto the dome?

I think what you might want is to put double illuminate on the dome so
that you can see it from the outside...

Post a reply to this message

Attachments:
Download 'dome_mod.zip' (57 KB)

"Paul Bourke" <pau### [at] uwaeduau> wrote in message 
news:web.48678631845cb4247e3508d40@news.povray.org...
> imagine a dome with 5 projectors creating an
> overlapping video image around the rim. See
>     http://local.wasp.uwa.edu.au/~pbourke/tmp/sim/
> I can convey how the images appear (see the model.zip file), but it breaks 
> down
> if I want to look at things from the outside of the dome.

How do you mean breaks down? Your model renders and the images are shown 
overlapping on the dome when viewed from outside the dome. Do you mean that 
the images are a bit dim? If so, turning up the light levels in the 
projectors to an artificial level such as 2 or 4 gives a brighter result.

Regards,
Chris B.

Post a reply to this message

> How do you mean breaks down? Your model renders and the images are shown
> overlapping on the dome when viewed from outside the dome.

Not so.

The parallel projection doesn't count, you can't tell the front face isn't being
illuminated due to symmetry.

The 90 degree perspective view is from the inside of the dome, 1 mm from the
back surface. If I move the camera back the image is not visible on the closest
face of the dome.

The spherical projection is obviously from the inside.

Post a reply to this message

> I think what you might want is to put double illuminate on the dome so
> that you can see it from the outside...

No, that is not the solution/problem.

I guess I need to learn how to drive photons, is there a good tutorial out
there?

Post a reply to this message

"Paul Bourke" <pau### [at] uwaeduau> wrote in message 
news:web.486a05c7945ca1b097a038180@news.povray.org...
>> How do you mean breaks down? Your model renders and the images are shown
>> overlapping on the dome when viewed from outside the dome.
>
> Not so.
>

Well I had thought that your camera #3 "side view" was from the outside, but 
it wasn't.

I added a camera #4 that is outside:
  camera {
     location -9000*z
     look_at 0
  }
and I can now see the problem that you were describing.

The following scene file illustrates a very different approach, so I don't 
know if it will serve your purposes at all. If so you'll need to adjust the 
'projector' positions and the scene file orientation, scale etc. The 
brightness of the overlaps won't be technically accurate (although you might 
be able to average the pigments) and it won't emulate optical components in 
the projectors (like mirrors), but it does solve the problem of the images 
being visible from both the inside and the outside.

Otherwise photons may be able to do what you want. I've not yet had much 
success with photons myself, so someone else would need to comment on 
whether they can be used to model how the light shining from a projector 
through a translucent image would fall on a screen.

Regards,
Chris B.

camera {location <0,0,-4> look_at 0 }
light_source {<-6,0 ,-4>, rgb 1}

sphere {
  0,1
  #declare I = 0;
  #while (I<5)
    texture {
      pigment {
        image_map {tga "projectorimage1.tga"
          once
          filter all 0.5
        }
        translate -<0.5,0.5,0>
        scale 0.2
        translate <0.5,0.5,0>
        warp {
          spherical
          orientation z
        }
      }
      // Projector Angle
      rotate -z*25
      // Projector Location
      translate <0.2,-0.6,0>
      rotate -90*y+I*360*y/5
      finish {ambient 0.8}
    }
    #declare I = I + 1;
  #end
  texture {
    pigment {image_map {tga "polargrid.tga"
        map_type 1
        interpolate 2
        once
      }
    }
  }
}

Post a reply to this message

Chris B wrote:

> Otherwise photons may be able to do what you want. I've not yet had much 
> success with photons myself, so someone else would need to comment on 
> whether they can be used to model how the light shining from a projector 
> through a translucent image would fall on a screen.

Well, I once tried this just to see if it worked and failed,
but succeeded now after revisiting the scene ;) At least, I
got *some* image projected onto the screen, which is rather
distorted but possibly correct for the crude "setup".

For realistic projections, I think the diameter of the light
source is also important. Shooting photons from a point light
has interesting effects, such as also getting an image when
no lens is present (simply magnified version of object).

Post a reply to this message

Attachments:
Download 'us-ascii' (3 KB)

> ... and it won't emulate optical components in the projectors (like mirrors),

Unfortunately that is exactly what I need, the resulting more complicated optics
(curved mirrors not just flat) is what I was hoping to emulate. Thanks anyway,
your "faked" (sorry it sounds harsh) was cunning anyway.

Post a reply to this message