Is there a reason why the calculation of trace intersections and their
normals in megapov may be less accurate than traces on CSG surfaces?

If so, is there a way to improve them:  super tight accuracy intervals,
etc.,?

I have been working on a golf game which involves bouncing a bunch of
particles off of an isosurface. For now, it kind of works like taking
200 drives from the tee and seeing if any of them fall into a hole far
away, after bouncing across an isosurface--a golf course.  I have spent
weeks pouring over my algorithm, and I am now convinced it is perfect.
The problem is that I am finding that a fraction of the "balls" fall
through the surface.  As I said, I am confident that my algorithm is
correct for determining intersections with the surface and the resultant
vector, even if it "bounces" 100 times during a "frame"!

However, some of my particles are still falling through the isosurface.
Consider:

CSG derived flat surface y=136         0 of 200 balls fall through
Isosurface derived flat surface y=136  7 of 200 balls fall through
Noise3d-involved bumpy isosurface     13 of 200 balls fall through

Is there something about the way the trace function interacts with
isosurfaces that I could use to fix this?

If I were simply making art, there would be no complaint. But I'm
considering making a "golf" competition, and thus someone might be upset
if there were a 6.5% chance of my algorithm being too stupid to keep the
ball from going underground! Some balls actualy "come back" after going
through with the bumpy surface!!

I can show people the whole file if you're interested in  a
disfunctional program.  But here's my macro at least:

#macro Bouncer
(rawvectin,resolvedvectin,posin,vecttestout,vecttrajout,posout,haddabounce)

        //rawvectin      is the instantaneous velocity, say  like 20 mph

        //resolvedvectin is the distance the particle will travel in one
frame's timespan, like say 2 inches
        //posin          is the particle position put into the equation.

        //vecttestout    is the distance the particle can still travel,
the "output" of resolvedvectin
        //vecttrajout    is the 'output' instantaneous velocity
        //posout         is the 'output' particle position
        //haddabounce    is 0 if no bounce occurred, 1 if it did.
        //funk              is the separately-called isosurface or
course.


        #declare Norm=<0,0,0>;
        #declare
crashpoint=trace(funk,posin,0.01*vnormalize(resolvedvectin),Norm);

        #if (Norm.x = 0 & Norm.y = 0 & Norm.z = 0)
                #declare posout=posin+resolvedvectin;
                #declare haddabounce=0;
                #declare vecttrajout=rawvectin-0.25*y;
        #else
                #declare
crashpct=1-vlength(posin-crashpoint)/vlength(resolvedvectin);
                // crashpct <0 if no crash during this frame
                // 0<crashpct<1 gives the fraction of resolvedvectin
that would be used up before hitting surface.

                #if(crashpct<0)
                        #declare posout=posin+resolvedvectin;
                        #declare haddabounce=0;
                        #declare vecttrajout=rawvectin-0.25*y;
                #else
                        #declare haddabounce=1;
                        #declare posout=crashpoint;//+.01*y;
                        #declare
vecttrajout=1.0*(rawvectin-2*vdot(rawvectin,Norm)*Norm);
                        #declare
vecttestout=crashpct*vlength(resolvedvectin)*vnormalize(vecttrajout);
                #end

        #end
#end

Post a reply to this message

Greg M. Johnson <gre### [at] my-dejanewscom> wrote:
: Is there a reason why the calculation of trace intersections and their
: normals in megapov may be less accurate than traces on CSG surfaces?

  Yes, there is a reason: Isosurfaces can't be calculated exactly, but
they have to be approximated. Increasing accuracy should lower the errors.

-- 
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):5;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/

Post a reply to this message