<abx### [at] babilonorg> wrote:

>On Thu, 06 Dec 2001 11:13:27 GMT, ken### [at] uniplanit (Angelo 'kENpEX' Pesce) wrote:
>> I'm only saying
>> that as I'm a softimage user, and many (many) other ppl use maya, or
>> 3dsmax (bleah!), povray should support them better
>
>But ... povray supports them better.
>Povray has well documented SDL which is independent platform for modelers.
>Moray works this way, many apps work this way.
>IMO it is not rule of pov-team to write good plugins for other apps to export
>scripts. IMO it is politic of high-expensive apps to not support _free_
>renderers. But it is personal opinion.

Well mabye this is the politic of high-expensive apps (dunno, I don't
see this politic, but mabye it is, or if it isn't it will be :P), and
I know that pov-team should not work to maya or softimage exporters.
But they could extend the SDL so that writing such exporters is easier
and comfortable...

>> > > Of course supporting both stuff will not hurt at all... :)
>> > And that's what POV has. Support for meshes and primitives. Make unlimited
>> > features with SDL or own patch and enjoy it.
>> As I already told U support for meshes is not enough... And I know
>> that I could make a patch and enjoy nurbs in povray, but as I have to
>> make a patch to enjoy nurbs in povray I'm saying here "please try to
>> support nurbs too"... :)
>Does 3ds (or other apps) support nurbs during tracing or generate triangles
>first and then perform rendering ? Why pov should support nurbs when modelers
>has implemented conversion from nurbs to triangles ? I don't want to say it
>shouldn't. I just think there are more important problems belonging to tracing.
>With good paper I can implement nurbs via macros. With functions{} I have nearly
>shader. I can't improve radiosity or media with macros or functions.

Well I can't tell that as many 3d gfx packages are coupled with a
default raytracer so it's difficoult to say when or where the modeller
ends and the raytracer starts... Ah yes, the thing that I can say to
you is that renderman scene language does support those primitives
natively, so as many renderers use that language they have to support
this stuff natively. I would like to tell U if mentalray does the same
too but I don't know (I don't think so as U can alter tessellation in
the editor, but who knows, mabye it passes the tessellation parameters
to the renderer...) So after all I still think, "yes it should support
them" :P

Update: I've found that mental ray supports nurbs too:
2.2. Curve and surface geometry





and arbitrary basis matrix, Taylor; rational and non-rational

any number of trimming curves, hole curves, and special curves and
points (that control tessellation) can be applied to free-form
surfaces

maximum free-form curve and surface degree of 21

adjacency detection (automatic detection and merging of neighboring
surfaces)

edge merging (cracks in inaccurate models are ``healed'')

curve approximation: 
two parametric modes 
edge length limit 
analytical distance limit 
angle limit 
view-dependent: closer objects are tessellated with more detail 

surface approximations: 
two parametric modes 
edge length limit 
analytical distance limit 
angle limit 
view-dependent: closer objects are tessellated with more detail 
Delaunay (can be combined with the above modes) 
graded Delaunay (can be combined with the above modes) 


quads (Catmull-Clark scheme) 
triangles (Loop scheme) 
wavelet-based multiresolution editing and data compression/LOD 
local refinement (per-face precision assignment) 
texturing 
sharp features: creases and points 
conversion from and to NURBS 
C++ modeling API 
displacement mapping: 
separate refinement approximations: parametric, edge length, distance,
angle; view dependency 
only small portions of a displacement-mapped object are tessellated as
needed. This allows very precise sub-pixel displacement that resolves
even the finest detail of the displacement map, without excessive
memory consumption 

So as U can see it's not the modelling program that should do it...
Another example? Lightflow, bmrt...

Are U still sceptical about this??
>ABX
>--
>#declare _=function(a,b,x){((a^2)+(b^2))^.5-x}#default {pigment{color rgb 1}}
>union{plane{y,-3}plane{-x,-3}finish{reflection 1 ambient 0}}isosurface{ //ABX
>function{_(x-2,y,1)|_((x+y)*.7,z,.1)|_((x+y+2)*.7,z,.1)|_(x/2+y*.8+1.5,z,.1)}
>contained_by{box{<0,-3,-.1>,<3,0,.1>}}translate z*15finish{ambient 1}}//POV35

Post a reply to this message