[Sidefx-houdini-list] basic instancing question

Olex P hoknamahn at gmail.com
Sun Oct 26 14:06:46 EDT 2008


Yes you can. Point is completely tha same primitive as polygon with only
one... two exceptions:
1. Shader executed once on point (i.e. it's impossible to get spherical like
shading on the single point)
2. You can't (I dunno why because it still make sense) displace point. As
workaround you can displace them in SOPs.
The rest is the same.

On Sun, Oct 26, 2008 at 4:56 PM, Pablo Giménez <pablogipi at gmail.com> wrote:

> 2008/10/26 Olex P <hoknamahn at gmail.com>
>
> > You could render them as Mantra points (using phase functions for
> > anisotropic scattering if you need it).
> > This could give you:
> > 1. Simpler setup
> > 2. Good details
> > 3. Super fast rendering
>
> Hey Olex when you render as Mantra points, can you apply a shader to them?
> Or the points only use the Cf and Of point attributes, I remember to have
> had some problems applying shaders to points in the pst, not sure now ...
>
> >
> >
> > On Sat, Oct 25, 2008 at 8:55 PM, jon parker <parker.jon at gmail.com>
> wrote:
> >
> > > Thanks for the repsonse...
> > > A colleague also tipped me to using displacement mapping to add detail
> > > to volumes, and it works really well.  For general volumetric
> > > rendering I'm in agreement with all of the things below.  With what I
> > > have been tasked with, there is a large scale difference in the nebula
> > > as a whole, versus the details we're trying to show... there are a
> > > large number of tiny 'knots' in the nebula that need to be in the
> > > right place according to some astronomical observation data I've been
> > > given.  My first attempt was to birth metaballs from the right
> > > locations and render them volumetrically, embedded within the rest of
> > > the gas volume.  But the shading quality of metaballs doesn't seem to
> > > be as nice as I hoped.  In fact I ran a test where I converted them to
> > > a massive image3d texture, and the shading quality was much nicer, but
> > > generating it took forever and that really hinders that technique for
> > > testing.  So my latest idea is to use a lot of knot-shaped volume
> > > instances, and displace them in world space to add variety, but they
> > > seem to be displaced before they are instanced, and not afterwards
> > > like I hoped.  Any other suggestions towards creating them this way?
> > > --jon
> > >
> > > On Sat, Oct 25, 2008 at 9:01 AM, Pablo Giménez <pablogipi at gmail.com>
> > > wrote:
> > > > Sounds pretty good.
> > > > The dispalcement to do the convection is quite interesting
> > > >
> > > > 2008/10/24 Remba, David (Space Technology) <David.Remba at ngc.com>
> > > >
> > > >>
> > > >> I've been working on clouds, and had good results
> > > >> using simple geometry (spheres, boxed shapes) then
> > > >> subdividing to round the shapes, then iso-offset to
> > > >> create a low density volume (20x20x20).
> > > >>
> > > >> The good stuff happens in a custom material:
> > > >>
> > > >> 1. displacement VOP SHOP uses various noise functions
> > > >> to push P around (for clouds: only in the up +Y direction
> > > >> to simulate convection, and using an abs() to make it
> > > >> puffy looking).  For nebula I would try the built-in
> > > >> aanoise VOP.
> > > >>
> > > >> 2. surface VOP SHOP based from the VEX Volumecloud shader,
> > > >> and added extra features for clouds to cut off and
> > > >> darken the bottom.  Higher density makes the gas more
> > > >> chunky looking, lower values make it more wispy looking
> > > >> (good for nebulas).
> > > >>
> > > >> tips:
> > > >>
> > > >> Very important to adjust the Volume Step size to as large
> > > >> a value as possible (just below the noticeable speckle artifact).
> > > >>
> > > >> For fast rendering: try only one or two lights; if you need
> > > >> more, do multiple render passes and comp them.  Use Ray Trace
> > > >> Engine, lower opacity limit to 0.9 if the nebula/clouds are
> > > >> chunky.
> > > >>
> > > >> For nebula you might try some glow effects (see the emit
> > > >> in the VEX Super Material shader).
> > > >>
> > > >> p.s. For nebulas, you might try separating very far objects
> > > >> which are more of a 2D problem, from closer 3D objects.
> > > >>
> > > >> p.p.s. I work at a space technology company so I might get
> > > >> to work on nebulas soon also!  I'd like to hear you things
> > > >> go for you.  My email:  daveremba at yahoo.com
> > > >>
> > > >> I hope this helps,
> > > >>
> > > >> Dave
> > > >>
> > > >> _______________________________________
> > > >> David Remba
> > > >> Tel 310 813 5899 Cel 310 809 6375
> > > >> Northrop Grumman Space Technology
> > > >> IMAGE Center
> > > >>
> > > >>
> > > >>
> > > >> -----Original Message-----
> > > >> From: sidefx-houdini-list-bounces at sidefx.com on behalf of jon
> parker
> > > >> Sent: Wed 10/22/2008 9:13 PM
> > > >> To: sidefx-houdini-list at sidefx.com
> > > >> Subject: [Sidefx-houdini-list] basic instancing question
> > > >>
> > > >> Greetings esteemed Houdini masters!
> > > >> I've been tasked to create a planetary nebula by some very demanding
> > > >> scientists... I've done some lighting for studios using Houdini in
> the
> > > >> past, but now the holes in my knowledge about Houdini are showing
> up,
> > > >> so I am looking for some basic workflow questions to be cleared up.
> > > >> For the nebula, the basic gas volume is looking great, however there
> > > >> are a large number of 'knots' in the nebula that need to be quite
> > > >> detailed.  I tried using volumetric metaballs first, but these don't
> > > >> seem to shade as nicely as the volume primitives... the density
> > > >> gradients don't seem as smooth, and the edge glows just aren't there
> > > >> like they are with volumes.  Converting them to i3d files gave me a
> > > >> much better look, but that's quite time consuming during look
> > > >> development, since I need at least 1024^3 i3d files to maintain the
> > > >> detail.  So I've tried instancing volumes to each knot, and this
> > > >> potentially looks best, except I have two questions about it...
> > > >> firstly, what is the best way to transfer point attributes to the
> > > >> volume?  Secondly, how can I use world space displacement on an
> > > >> instance?  On the first attempt I discovered that instances
> themselves
> > > >> are not displaced; instead they pick up the displacement of the
> source
> > > >> geo before they are instanced.   Aside from this question, which
> > > >> approach makes the most sense?  There are obviously a lot of
> potential
> > > >> paths to take in Houdini.
> > > >> Cheers,
> > > >> Jon
> > > >> _______________________________________________
> > > >> Sidefx-houdini-list mailing list
> > > >> Sidefx-houdini-list at sidefx.com
> > > >> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > > >>
> > > >>
> > > >> _______________________________________________
> > > >> Sidefx-houdini-list mailing list
> > > >> Sidefx-houdini-list at sidefx.com
> > > >> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > > >>
> > > >>
> > > >
> > > >
> > > > --
> > > > Un saludo
> > > > Best Regards
> > > > Pablo Giménez
> > > > _______________________________________________
> > > > Sidefx-houdini-list mailing list
> > > > Sidefx-houdini-list at sidefx.com
> > > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > > >
> > > _______________________________________________
> > > Sidefx-houdini-list mailing list
> > > Sidefx-houdini-list at sidefx.com
> > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > >
> > _______________________________________________
> > Sidefx-houdini-list mailing list
> > Sidefx-houdini-list at sidefx.com
> > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >
>
>
>
> --
> Un saludo
> Best Regards
> Pablo Giménez
> _______________________________________________
> Sidefx-houdini-list mailing list
> Sidefx-houdini-list at sidefx.com
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>



More information about the Sidefx-houdini-list mailing list