On Wed, Nov 29, 2000 at 08:26:47AM -0500, Chris Pelkie wrote:
> First, be aware that many modules, including AutoGlyph have "hidden"
> parameters. Open the configuration dialog box by double-clicking the
> module, then Expand to see these. That said, I don't think that any
> permutation of ratio/min/max will eliminate the display of the spheres
> corresponding to a mag(vector)=0.0; they just change which vectors are
> rendered as spheres.
> 
> So, unless someone else provides an easier way, I would use Include to mask
> these off. Include can either cull them out or you can create invalid
> positions to mask them. I'm proposing to do the latter.
> 
> If the vector data is the "data" component of field F, use
> mags = Compute(F, "mag(a)!=0.0"); //(or "mag(a)==0.0 ? 0 : 1")
> invs = Include(mags, min=0.5, cull=0);
> // you don't want cull=1 (the default) because you need a component
> // of "invalid positions" of the same count as the original vector array
> // to use in the next step
> maskedF = Replace(invs, F, "invalid positions", "invalid positions");
> 
> Result is that every mag(a)=0 is now masked by an "invalid position" of 01,
> so it won't render (they are still in the maskedF field however in case you
> want to do something else with them).
> 

Hi Chris,

Excellent, thanks, it's just perfect, I'll do this.
Just yesterday I learned about the Include module, I used it to
make one of those stunning pictures where you chop off a part of
a 3D finite element mesh (I'm working with CFD) and see part of the
surface mesh of the body stuck in a `block' formed by the remaining
elements. Heck, I hope you understood what I meant because I can't
explain it with words but it's breathtaking.
I love DX, people here at my department think I'm a god!

Regarding you comment to the other suggestion, I'm also very interested
in memory optmization, even though the meshes I'm working with are
a lot smaller than yours (about 60000 nodes, 250000 elements). It's
funny to see how quicky and gladly DX takes up 200MBytes. I still have some
left, but in the near future intend to work with larger meshes, so memory
is an issue to me.

Many thanks again,

-- 
Ramiro Brito Willmersdorf            [EMAIL PROTECTED]  
GPG key: http://www.demec.ufpe.br/~rbw/GPG/gpg_key.txt

Reply via email to