Le_Forgeron <jgr### [at] freefr> wrote:

>
> It make sense, as for the old airplane's cockpit, only if there is a
> third guy on the back, maybe turned sideway.

Yes, like a flight engineer. Of course, this goofy police car is probably
completely computerized and 'fly-by-wire'-- so gauges and dials would be
redundant. But they *look* important. ;-)

BTW, the photo I used *is* from an old WWII airplane cockpit. In fact, I've used
it before, for the interior of my B-29 bomber model (posted a couple of years
ago.)

Post a reply to this message

On Sat, 04 May 2013 00:52:24 +0200, Kenneth <kdw### [at] gmailcom> wrote:


> BTW (and this is a bit off-topic): For your jazz-musician scene and its
> neon-signs (the sphere_sweeps), did you experience any of the 'clipping'
> problems that are being discussed? I didn't see any obvious problems in  
> your
> renders.
>

I really don't remember having any clipping problems. It was hand-coded.

-- 
-Nekar Xenos-

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:
> The spline requirement for the sphere_sweep is indeed a tricky thing to come up
> with (re: my particular windshield outline.) What I imagine is that the spline
> might not end up exactly following the correct curve. But that's just
> theorizing; I should give it a chance, to see what happens.

You could always use a linear spline with lots of points, this would follow your
spline points exactly and can be made to look pretty smooth e.g.

union
{
sphere_sweep {
  linear_spline                 // linear curve
  //cubic_spline                // alternative spline curves
  //b_spline
  #declare np=80;
  #declare p=0;
#declare r=0;
  np+1,                            // number of specified sphere positions
#while ( p < np+1 )
  vrotate(<0,-10, 0>, <0, 0, r>), 1
  #declare r=r+360/np;
  #declare p=p+1;
  #end
 // tolerance 0.001             // optional

  pigment { rgb <1, 0, 0> }
  finish { specular 0.5 roughness 0.01 }
}
  bounded_by { box { <-11, -11, -1> <11, 11, 1> } }
 }


produces the torus shape below. I recommend adding a bounded_by as performance
is pretty poor otherwise and probably as part of the spline bounding issue it
seems that you have to put the sphere_sweep inside a union and bound the union
otherwise it makes no difference to performance.

Sean

Post a reply to this message

Attachments:
Download 'sphere_sweep.bmp.dat' (1407 KB)

"s.day" <s.d### [at] uelacuk> wrote:

>
> You could always use a linear spline with lots of points, this would
> follow your spline points exactly and can be made to look pretty
> smooth...
>

That idea didn't occur to me. (And I've never noticed *any* problems with the
linear spline.) I can see this working quite well (as your example image
demonstrates; thanks for that.) In fact, the 'linear' nature of the final
shape--if looked at on a tiny scale-- would not be too far removed from what I'm
doing now. That is, my 3000X3000 preliminary image render of the windshield
area, when seen close-up, has a 'pixel stairstep' appearance anyway. In essence,
linear--and it's being scanned that way too. EXCEPT that a sphere_sweep would be
*smooth*-- no 'micro bumpiness' like my current many-spheres technique.

This is an intriguing idea! The thing is, I've forgotten if a sphere_sweep
object is 'evaluated' by POV-Ray at parse time or during the rendering stage. If
the latter, that might slow things down during my animation (regardless of
#writing the sphere_sweep to a file beforehand.) And maybe the same situation
using blobs.

I'll have to try this to see.


> I recommend adding a bounded_by as performance
> is pretty poor otherwise and probably as part of the spline bounding issue it
> seems that you have to put the sphere_sweep inside a union and bound the union
> otherwise it makes no difference to performance.

Thanks for the good advice!

Post a reply to this message

On Sun, 05 May 2013 18:48:17 +0200, Kenneth <kdw### [at] gmailcom> wrote:

> This is an intriguing idea! The thing is, I've forgotten if a  
> sphere_sweep
> object is 'evaluated' by POV-Ray at parse time or during the rendering  
> stage. If
> the latter, that might slow things down during my animation (regardless  
> of
> #writing the sphere_sweep to a file beforehand.) And maybe the same  
> situation
> using blobs.
>
> I'll have to try this to see.
>
>

Apparently blobs render faster:
http://news.povray.org/povray.windows/thread/%3Cweb.47841f263bacf7e1748549bc0@news.povray.org%3E/?ttop=381969&toff=50

-- 
-Nekar Xenos-

Post a reply to this message

Have you seen the old blob-modeled work by H.E. Day? Similar to this. His
absorbing/emitting media code may still be online somewhere. If not, the
technique is well-documented.

 -Shay

Post a reply to this message