> 
>   That doesn't mean POV-Ray will never ever implement such algorithms if
> they are good.
> 

I would love to see the code (3.7 or more probably 4.0) abstract things 
enough so that adding objects, camera projections, texture patterns, or 
*whole rendering/lighting algorithms* to be added without having to 
touch too much core code. For example, adding a new rendering algorithm 
that is a forwards-raytracer that would work with all existing objects, 
cameras, patterns, pigments, and normals; only having to change the 
finish and maybe the light_source for the lighting features specific to 
the algorithm.

I saw a quite interesting renderer using forwards raytracing (I'm not 
sure if it's completely unbiased, looks good enough), but doesn't even 
have *CSG* on its scene description language. I want CSG, isosurfaces 
and a spherical camera; with *that* lighting.

Post a reply to this message

> that would work with all existing objects, 
> cameras, patterns, pigments, and normals; only having to change the 
> finish and maybe the light_source for the lighting features specific to 
> the algorithm.

Correction: oops, renderers that don't use the normal vector (like 
WinOSi[1]) won't work with normal{}.

[1] Note 2 in http://www.winosi.onlinehome.de/Docs.htm

Post a reply to this message

"Warp" <war### [at] tagpovrayorg> wrote in message
news:4721ca0a@news.povray.org...

>   POV-Ray is the unfortunate victim of negative hype. One reason for this
> is that POV-Ray has existed for so long and doesn't jump to every trend
> which is being currently fashionable and hyped.

Neither quality nor speed (both set-up and rendering) is hype, it's the
reality and business.

Post a reply to this message

somebody wrote:
...
> ...B-rep modelling, whether mesh or
> NURBS is also a must, neither animation nor manufacture is easy or natural
> with implicit "modelling" as in SDL, which only made sense 20 years ago when
> realtime b-rep modelling wasn't feasible on a budget.
...

It IS possible to model with NURBS in POV-Ray, by using SDL.

BUT it is very hard to model with it without having the
possibility to move the control points around (or changing
the weights or the knot vectors) with a pointing device
like e.g. a mouse.

-- 
Tor Olav
http://subcube.com

Post a reply to this message

"Tor Olav Kristensen" wrote:
> It IS possible to model with NURBS in POV-Ray, by using SDL.




splines, which can be regarded as uniform, non-rational B-splines.



Rune

Post a reply to this message

"Rune" <aut### [at] runevisioncom> wrote:
> "Tor Olav Kristensen" wrote:
> > It IS possible to model with NURBS in POV-Ray, by using SDL.
>

>

> splines, which can be regarded as uniform, non-rational B-splines.
>

>
> Rune

http://news.povray.org/povray.binaries.images/thread/%3CXns935A4D394502torolavkhotmailcom%40204.213.191.226%3E/?mtop=10


Tom

Post a reply to this message

"Tom York" <alp### [at] zubenelgenubi34spcom> wrote in message 
news:web.4722362c8d03a40e7d55e4a40@news.povray.org...
> "Rune" <aut### [at] runevisioncom> wrote:
>> "Tor Olav Kristensen" wrote:
>> > It IS possible to model with NURBS in POV-Ray, by using SDL.
>>

>>

>> splines, which can be regarded as uniform, non-rational B-splines.
>>

>
>
http://news.povray.org/povray.binaries.images/thread/%3CXns935A4D394502torolavkhotmailcom%40204.213.191.226%3E/?mtop=10

Ah okay, I misunderstood the "by SDL" part in Tor's message. I thought he 
meant modeling NURBS by SDL, not implementing NURBS by SDL.

Rune

Post a reply to this message

"Tor Olav Kristensen" <tor### [at] TOBEREMOVEDgmailcom> wrote
> somebody wrote:

> > ...B-rep modelling, whether mesh or
> > NURBS is also a must, neither animation nor manufacture is easy or
natural
> > with implicit "modelling" as in SDL, which only made sense 20 years ago
when
> > realtime b-rep modelling wasn't feasible on a budget.

> It IS possible to model with NURBS in POV-Ray, by using SDL.
>
> BUT it is very hard to model with it without having the
> possibility to move the control points around (or changing
> the weights or the knot vectors) with a pointing device
> like e.g. a mouse.

Well, that's the point. It is of course possible to model *any* mesh or
NURBS based model within the SDL, after all, I am pretty sure that the SDL
is equivalent to a universal Turing machine. You could, if all else fails,
place ten million triangles by hand. That doesn't mean it's desirable or
feasible to do so. Nobody that I know of models NURBS surfaces by typing
down control points, knot vectors, weights and trimming profiles by hand in
a production environment.

Post a reply to this message

delle wrote:
> Take a look to Indigo Render Gallery:

Color me unimpressed. Maybe it's because I'm not an expert, but some of 
the sub-surface scattering stuff is the only stuff that looks 
particularly good to me. Balanced against most of their proud gallery 
being obnoxiously grainy, I don't see it as a win just from the photos.

Is it possible to automatically know when a scene is good enough? Or 
does it take human intervention to say "ok, stop now and move on to the 
next frame"?

-- 
   Darren New / San Diego, CA, USA (PST)
     Remember the good old days, when we
     used to complain about cryptography
     being export-restricted?

Post a reply to this message

> delle wrote:
>> Take a look to Indigo Render Gallery:
>
> Color me unimpressed. Maybe it's because I'm not an expert, but some of 
> the sub-surface scattering stuff is the only stuff that looks particularly 
> good to me. Balanced against most of their proud gallery being obnoxiously 
> grainy, I don't see it as a win just from the photos.
>
> Is it possible to automatically know when a scene is good enough? Or does 
> it take human intervention to say "ok, stop now and move on to the next 
> frame"?

--- Indigo's Gallery ---
There is a lot of beautiful art there, no doubt.  The modeling of the
scenes is particularly good in most cases. I'm not totally convinced
at the quality of Indigo though...

Many images there are grainy, or contain grainy spots.

Many of the global illumination (radiosity) scenes had posted times
over 12 hours, when I'm sure equivalent scenes in POV would
render in less time (usually) and not be grainy.

Only a few of the glassware images even tried to reproduce any
IOR or photons, and looked particularly fake in most cases.

--- NVIDIA's Gelato Gallery ---
There's some pretty amazing scenes here that have a huge amount
of geometry.  If I wanted to trace a billion triangles I'm sure NVIDIA
cards could do it in some manner.

The global illumination here seems to suffer from being grainy,
like Indigo, maybe a little grainier.

Glassware is much better than Indigo's, good photons,
somewhat blocky edges at times.

The subsurface scattering looks nice (we should put this into POV).

Many of the images seemed saturated.

--- Maxwell's Gallery ---
Just stunning, very good in most every respect.

I did notice just a couple grainy spots, but you needed to look for
that, mostly it's not obvious.

Of course setting this package up with their modeler will set
you back about 3400 pounds... $7000

--- Maya / Mental Ray / AutoCad ---
Yes, Autodesk is now selling Maya with Mental Ray licenses.
Equivalent or better than POV in quality, with a proven
track record.

Post a reply to this message