Hi,

does anyone have an answer to help me out here ?

I am trying to render an animation of the surface of a biological
molecule, which is composed of thousands of triangles.
The rendering of such an animation takes a few minutes per frame. 

I would now like to cut the molecule half-open to see the interior,
while it rotates. If I introduce a 

command, the rendering time decays to about 12 hours per frame.
The same slowdown happens, when I use the difference command.

Does anybody know how to speed this up ?

Thanks,

Hen### [at] unibasch

Post a reply to this message

mihm17 wrote:
> 
> Hi,
> 
> does anyone have an answer to help me out here ?
> 
> I am trying to render an animation of the surface of a biological
> molecule, which is composed of thousands of triangles.
> The rendering of such an animation takes a few minutes per frame.
> 
> I would now like to cut the molecule half-open to see the interior,
> while it rotates. If I introduce a
> clipped_by { box { <-1000,-1000,-1000>, <1000,0,1000> }
> command, the rendering time decays to about 12 hours per frame.
> The same slowdown happens, when I use the difference command.
> 
> Does anybody know how to speed this up ?

The reason this is taking so long is because you are doing a csg
operation on 1000's of objects. Each object that is intersected
by the box must be evaluated seperately which greatly increases
the rendering time. You may be able to speed things us a bit by
manually bounding the object(s) in question but I doubt the increase
will be signifigant. If you can figure out a way to create your
molecule object out of a few primitive objects instead of serveral
thousand traingles you can probably speed things up quite a bit.

-- 
Ken Tyler - 1400+ POV-Ray, Graphics, 3D Rendering, and Raytracing Links:
http://home.pacbell.net/tylereng/index.html http://www.povray.org/links/

Post a reply to this message

Thanks Ken.

I cannot reduce the number of triangles my scenery is composed of.
The biological molecule is that complicated, and the electron-microscopical

3D structure of that membrane-protein should be scientifically visualized
without simplifications or alterations.

I used now a
Pre_Frame_Commmand = my_clip %k
and clip the set of triangles outside of POVray. This works, and I can
calculate again
a frame in a few minutes.

BUT: Why is this faster than POVray can do ?
Is it that POVray is clipping the whole 3D scenery for every pixel/ray all
over again ?

Wouldn't it be much faster to once clip the 3D scenery while reading the
scene,
and then do the raytracing without further clipping calculations ?

Or is there a kind of Preprocessor incorporated into POVray ?

Yours, Henning.


Ken wrote:

> mihm17 wrote:
> >
> > Hi,
> >
> > does anyone have an answer to help me out here ?
> >
> > I am trying to render an animation of the surface of a biological
> > molecule, which is composed of thousands of triangles.
> > The rendering of such an animation takes a few minutes per frame.
> >
> > I would now like to cut the molecule half-open to see the interior,
> > while it rotates. If I introduce a
> > clipped_by { box { <-1000,-1000,-1000>, <1000,0,1000> }
> > command, the rendering time decays to about 12 hours per frame.
> > The same slowdown happens, when I use the difference command.
> >
> > Does anybody know how to speed this up ?
>
> The reason this is taking so long is because you are doing a csg
> operation on 1000's of objects. Each object that is intersected
> by the box must be evaluated seperately which greatly increases
> the rendering time. You may be able to speed things us a bit by
> manually bounding the object(s) in question but I doubt the increase
> will be signifigant. If you can figure out a way to create your
> molecule object out of a few primitive objects instead of serveral
> thousand traingles you can probably speed things up quite a bit.
>
> --
> Ken Tyler - 1400+ POV-Ray, Graphics, 3D Rendering, and Raytracing Links:
> http://home.pacbell.net/tylereng/index.html http://www.povray.org/links/

Post a reply to this message

In article <3A6D57E3.F19A6457@unibas.ch>, hen### [at] unibasch 
wrote:

> BUT: Why is this faster than POVray can do ?
> Is it that POVray is clipping the whole 3D scenery for every pixel/ray all
> over again ?
> 
> Wouldn't it be much faster to once clip the 3D scenery while reading 
> the scene, and then do the raytracing without further clipping 
> calculations ?

POV can't do that, because it can't know that an intersection won't be 
used until it computes that intersection. The necessary information just 
doesn't exist until POV is tracing a pixel...

Hmm, I just thought of something...you may be expecting the scene to be 
converted into triangles, which would be discarded at parse time. 
POV-Ray traces most objects directly, solving the mathematical 
representations of the objects, not converting them to a mesh of 
triangles.

-- 
Christopher James Huff
Personal: chr### [at] maccom, http://homepage.mac.com/chrishuff/
TAG: chr### [at] tagpovrayorg, http://tag.povray.org/

<><

Post a reply to this message