Nicolas Alvarez wrote:
> Remember I can't use radiosity on the renderfarm due to the artifacts 
> that appear on tiled radiosity renders.

Question: How well does radiosity hold up across multiple frames if each 
frame can be made into one tile?

Though, for the animation I have in mind, a light dome would probably be 
more than adequate. Which means that I could spend the render time on 
scattering media instead! :-D

-- 
William Tracy
afi### [at] gmailcom -- wtr### [at] calpolyedu

This summer...There is another word for EXCITEMENT!
Roget's Thesaurus: The Motion Picture
     -- posted by dr_dank on Slashdot

Post a reply to this message

> Nicolas Alvarez wrote:
>> Remember I can't use radiosity on the renderfarm due to the artifacts 
>> that appear on tiled radiosity renders.
> 
> Question: How well does radiosity hold up across multiple frames if each 
> frame can be made into one tile?

Good enough, if POV-Ray versions match. I remember a terribly flickering 
animation the Internet Movie Project did, and the problem was some users 
had different POV versions. That's not a problem on my farm, since 
everyone is using the exact same version (and with my modifications).

The main problem is there *might* be artifacts when resuming a render. I 
can't guarantee frames will be rendered in "one go" (people may reboot, 
or BOINC automatically decide to switch from raytracing to protein folding).

Post a reply to this message

> I started with the untextured version (normally used for radiosity pass).
Here the same scene but with texture.

Any comments wellcome.

Regards,
Fidos

Post a reply to this message

Attachments:
Download '2nd-pass.03.jpg' (169 KB)

Preview of image '2nd-pass.03.jpg'

how come the bottle is black ?

cu!
-- 
#macro G(b,e)b+(e-b)*C/50#end#macro _(b,e,k,l)#local C=0;#while(C<50)
sphere{G(b,e)+3*z.1pigment{rgb G(k,l)}finish{ambient 1}}#local C=C+1;
#end#end _(y-x,y,x,x+y)_(y,-x-y,x+y,y)_(-x-y,-y,y,y+z)_(-y,y,y+z,x+y)
_(0x+y.5+y/2x)_(0x-y.5+y/2x)            // ZK http://www.povplace.com

Post a reply to this message

ah, it's not just the bottle, it's every reflective or refractive surface. 
or so it seems.

cu!
-- 
#macro G(b,e)b+(e-b)*C/50#end#macro _(b,e,k,l)#local C=0;#while(C<50)
sphere{G(b,e)+3*z.1pigment{rgb G(k,l)}finish{ambient 1}}#local C=C+1;
#end#end _(y-x,y,x,x+y)_(y,-x-y,x+y,y)_(-x-y,-y,y,y+z)_(-y,y,y+z,x+y)
_(0x+y.5+y/2x)_(0x-y.5+y/2x)            // ZK http://www.povplace.com

Post a reply to this message

"Zeger Knaepen" <zeg### [at] povplacecom> wrote:
> ah, it's not just the bottle, it's every reflective or refractive surface.
> or so it seems.

yes, the original scene had a table with glass tabletop.  It wasn't black, but
seemed to not show reflections or much transparency.  Perhaps should get a
higher max_trace/whatever count?

OTOH, I like where this is going.  In the first test, the metal parts didn't
have the jaggies in Jaime's original scene...

Post a reply to this message

BTW, these comparisons are cool?

http://luxrender2.org/forum/viewtopic.php?f=13&t=420

It's fascinating to see pictures with relatively low noise like this:
http://www.ossosso.com/public/test/studio_luxrender.tif

in just 5 minutes.

Not two mention no complicated tweaking to make materials and lights behave as
they should...

Post a reply to this message

"nemesis" <nam### [at] gmailcom> wrote:
> "Zeger Knaepen" <zeg### [at] povplacecom> wrote:
> > ah, it's not just the bottle, it's every reflective or refractive surface.
> > or so it seems.
>
> yes, the original scene had a table with glass tabletop.  It wasn't black, but
> seemed to not show reflections or much transparency.  Perhaps should get a
> higher max_trace/whatever count?
I need to use 2 different settings : one for the max recursion of a ray on the
diffuse objects and one the for the max trace level. An other possible reason
is that I perhaps break the refraction (reflection seems work)...

Post a reply to this message

fidos wrote:
>> I started with the untextured version (normally used for radiosity pass).
> Here the same scene but with texture.
> 
> Any comments wellcome.

I'm very impressed with the results! So much more natural lighting than
the original with Povray's radiosity - at least with the settings
Piqueres used here:

http://www.ignorancia.org/uploads/images/thekitchen/wips/thekitchen-29.jpg

I've been following the work with WinOSI but it seems that the
development is halted. And as it is written in assemply I don't see any
future in any case :)

But please keep on working with this patch as the results are so good!
Could you post a comparison renders from various scenes with hi quality
radiosity settings. And report the actual rendering time.

I also guess that it should be very easy to parallelize Monte Carlo path
 tracing to take advantage of multiple CPU cores.

Post a reply to this message

Warp wrote:

>   So you basically do the exact same thing that the current radiosity
> code does, but without the optimizations? (The current radiosity method
> *uses* the monte-carlo method, with optimizations.)
> 
>   I really fail to see the advantage.

I can at least see that there are big differences in the images posted
to this thread - of course, that doesn't necessarily mean anything. And
I like images done with Fidos's patch the most. It would be interesting
to see if one could achieve identical lighting with Povray's radiosity.
The stupid thing is that you have to tweak the settings so much - with
brute force you just...wait :)

Post a reply to this message