Greetings from Cologne! 

May be, the following reflections are not new !?
May be just unnecessary or pesky!?
In this case give me a piece of your mind !

I've rarely gotten satisfactory results using (Mega)POV's radiosity. 
Therefore, in my last projects, I add to almost all surfaces (except 
perhaps the glass) MegaPov's "blurred reflection" - that is, what happens 
in nature - it's true in opposite direction - but it's the habit of raytracing 
anyway. 

I'm quite happy with the results and it's easy to use. 
With the few settings you have good control over the appearence of each 
object and the results are correct, even in small details. 

For most surfaces I use the following settings: 
	reflection .5
	reflection_blur .5
	reflection_samples 50
	reflect_metallic 1
	and
	a bright hollow sphere or a sky_sphere to spend the diffuse light.

..
BUT, there are two faults, and they are the reasons for this posting: 

	1) The granulation
	2) The slight shift of the diffuse reflections towards the camera 

My suggestions are: 
	
	Ad 1) 
	I think, the granulation could be avoid by adding an option to the 
	"blurred reflection", which switches off the randomness of the 
	direction of the reflected rays in each point. 
	So - each point of surface should get the same, regular bunch of 
	rays. Perhaps such a bunch (reflection_samples) needs less 
	samples, which will reduce the rendertime as well. 
	
	Ad 2)
	The reflection shift could be avoid by using a blurred reflection 
	(bunch of reflected rays), always PERPENDICULAR TO THE 
	SURFACE, regardless of the visual angle between the surface 
	and camera point. 

I'm not experienced in changing and compiling source-code - 
so, perhaps, perhaps, someone...!? 

.. what shall I do? 


Thank you for your patience!
I'm curious about reactions - and, of course, Help!


Tom Aust
www.aust-manufaktur.de

Post a reply to this message

Attachments:
Download 'radtest.jpg' (99 KB)

Preview of image 'radtest.jpg'

Tom Aust wrote:
> 
> Greetings from Cologne!
> 
> May be, the following reflections are not new !?
> May be just unnecessary or pesky!?
> In this case give me a piece of your mind !
> 
> I've rarely gotten satisfactory results using (Mega)POV's radiosity.
> Therefore, in my last projects, I add to almost all surfaces (except
> perhaps the glass) MegaPov's "blurred reflection" - that is, what happens
> in nature - it's true in opposite direction - but it's the habit of raytracing
> anyway.
> 
> [...]

The picture looks fairly interesting, but that does not really have much
to do with radiosity in fact.  As you mentioned the concept of reflections
in raytracing is limited to one direction so you might want to try photons
too.

I think some time ago someone suggested in p.g. that blurred reflection
and photons combined would be sufficient to generate realistic surfaces,
but the necessary amount of computations would be enormous. 

Christoph

-- 
Christoph Hormann <chr### [at] gmxde>
IsoWood include, radiosity tutorial, TransSkin and other 
things on: http://www.schunter.etc.tu-bs.de/~chris/

Post a reply to this message

Tom Aust <aus### [at] aust-manufakturde> wrote:
: 	The reflection shift could be avoid by using a blurred reflection 
: 	(bunch of reflected rays), always PERPENDICULAR TO THE 
: 	SURFACE, regardless of the visual angle between the surface 
: 	and camera point. 

  So what you want to achieve is, in fact, the radiosity algorithm used
by POV-Ray, but a LOT slower (because you want to calculate ALL the reflected
rays for EACH intersection point).

-- 
#macro N(D,I)#if(I<6)cylinder{M()#local D[I]=div(D[I],104);M().5,2pigment{
rgb M()}}N(D,(D[I]>99?I:I+1))#end#end#macro M()<mod(D[I],13)-6,mod(div(D[I
],13),8)-3,10>#end blob{N(array[6]{11117333955,
7382340,3358,3900569407,970,4254934330},0)}//                     - Warp -

Post a reply to this message

"Tom Aust" <aus### [at] aust-manufakturde> wrote...

> So - each point of surface should get the same, regular bunch of
> rays. Perhaps such a bunch (reflection_samples) needs less
> samples, which will reduce the rendertime as well.

But this would cause artifacts (like when you use a 3x3 area_light without
jitter).

> The reflection shift could be avoid by using a blurred reflection
> (bunch of reflected rays), always PERPENDICULAR TO THE
> SURFACE, regardless of the visual angle between the surface
> and camera point.

You'll be happy to know that it's been done!
http://perso.wanadoo.fr/albedo/patchedpov.html

But you may want to look here:
http://perso.wanadoo.fr/albedo/cornell.html

to see what to expect for render times.

-Nathan

Post a reply to this message

Nathan Kopp wrote:
>
> You'll be happy to know that it's been done!
> http://perso.wanadoo.fr/albedo/patchedpov.html
> 
> But you may want to look here:
> http://perso.wanadoo.fr/albedo/cornell.html
> 
> to see what to expect for render times.

Those pov scenes use omnidirectional light sources so they really can't
be compared to the Cornell rendering. Here is a POV 3.5 pre-beta rendering
of the same scene geometry using a light source that matches the diffuse
patch used in the original. Render time 5 minutes on 1GHz system. 
http://www.pp.htv.fi/kkivisal/cornell_beta.jpg

Here is a blurred reflection version of the same scene. 30 minutes.
http://www.pp.htv.fi/kkivisal/cornell_blur.jpg


_____________
Kari Kivisalo

Post a reply to this message