Hi all:

As strange as it sounds, these two renderings use the exact same 
radiosity settings, and the same lighting (no light_sources, only a 
tricky image_map function onto a surrounding sphere).

The only difference is an extra texture layer for the lathe objetcs. It 
started as a dust layer, but accidentally resulted into an artifact cleaner:

texture{
  pigment{
   wrinkles
   color_map{
    [0.0 rgbt 1]
    [0.0 rgbt .9]
   }
  }
}

I've still to test if this works more generally, not only for lathe 
objetcs, but I wanted to share this strange but useful "effect". If I 
can isolate it I will post a minimal scene later.

--
Jaime

Post a reply to this message

Attachments:
Download 'rad-artifact-cleaner-0.jpg' (17 KB) Download 'rad-artifact-cleaner-1.jpg' (15 KB)

Preview of image 'rad-artifact-cleaner-0.jpg'

Preview of image 'rad-artifact-cleaner-1.jpg'

Jaime Vives Piqueres wrote:
> Hi all:
> 
> As strange as it sounds, these two renderings use the exact same 
> radiosity settings, and the same lighting (no light_sources, only a 
> tricky image_map function onto a surrounding sphere).
> 
> The only difference is an extra texture layer for the lathe objetcs. It 
> started as a dust layer, but accidentally resulted into an artifact 
> cleaner:
> 
> [...]

How do the render times differ?  The layered texture version is probably 
quite a bit slower since the lighting is evaluated for both layers.

BTW are you sure both color map indices are 0.0?

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 07 Mar. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

> How do the render times differ?  The layered texture version is probably 
> quite a bit slower since the lighting is evaluated for both layers.

   Only a tiny bit slower... and I was going to use an extra layer anyhow.


> BTW are you sure both color map indices are 0.0?

   Hmmm... yes and no. It was a typo on the scene code, not on the 
message. :(

   So, it means that texture{pigment{rgbt .9}} alone does shows this 
effect. I tested it a bit more, and it's not as useful as I happily 
assumed. It does weird things with plain surfaces, and only softens 
certain kind of artifacts, with somewhat better results over curved 
surfaces. And it introduces squared artifacts when loading rad data.

   The attached images show a simple radiosity scene with only 1 
recursion level and intentional big artifacts. The first one is with 
simple textures, the second uses an extra layer "texture{pigment{rgbt 
.95}}" on all objetcs, and the last uses the same extra layer but was 
rendered loading the radiosity data (code on p.t.scene-files).

   On the lathes I was doing, it worked great. But only were useful on 
the vases, not for the stand because it showed ugly squares and the 
artifacts were not as evident (I will post some examples on a new message).

--
Jaime

Post a reply to this message

Attachments:
Download 'rad-artifact-cleaner0.jpg' (25 KB) Download 'rad-artifact-cleaner1.jpg' (22 KB) Download 'rad-artifact-cleaner1l.jpg' (22 KB)

Preview of image 'rad-artifact-cleaner0.jpg'

Preview of image 'rad-artifact-cleaner1.jpg'

Preview of image 'rad-artifact-cleaner1l.jpg'

Jaime Vives Piqueres wrote:
> 
>   Hmmm... yes and no. It was a typo on the scene code, not on the 
> message. :(
> 
>   So, it means that texture{pigment{rgbt .9}} alone does shows this 
> effect. I tested it a bit more, and it's not as useful as I happily 
> assumed. It does weird things with plain surfaces, and only softens 
> certain kind of artifacts, with somewhat better results over curved 
> surfaces. And it introduces squared artifacts when loading rad data.
> 
>   The attached images show a simple radiosity scene with only 1 
> recursion level and intentional big artifacts. The first one is with 
> simple textures, the second uses an extra layer "texture{pigment{rgbt 
> .95}}" on all objetcs, and the last uses the same extra layer but was 
> rendered loading the radiosity data (code on p.t.scene-files).

Note that reloading radiosity data might make an important difference 
since not all data gets saved to the file (some of the irradiance 
gradient information is lost, this might be the reason for your squared 
artifacts).  Have you tried adding a normal modifier to the second 
texture layer (not much, just a tiny variation)?

I should also add that with high ambient surfaces you should make sure 
you use the 3.6 beta since there were some modifications in the 
radiosity code.  Since the overall lighting level of the scenes is not 
that high this probably does not make much difference but who knows...

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 07 Mar. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

> Note that reloading radiosity data might make an important difference 
> since not all data gets saved to the file (some of the irradiance 
> gradient information is lost, this might be the reason for your squared 
> artifacts).  Have you tried adding a normal modifier to the second 
> texture layer (not much, just a tiny variation)?
>
> I should also add that with high ambient surfaces you should make sure 
> you use the 3.6 beta since there were some modifications in the 
> radiosity code.  Since the overall lighting level of the scenes is not 
> that high this probably does not make much difference but who knows...

   Thanks! I combined your 2 suggestions in one (downloading the beta 
and using a normal), and the results are slighty better.

   The best result I achieved, was using the filtered layer with a 
little normal when saving rad data (with "normal on"), and later 
reloading the data without the normal. But it still has some problems, 
as even worse artifacts showing on reflections if these are not included 
when saving rad data.

   It can help a bit with poor/medium quality settings, and in fact I 
was using/suffering it for sometime without noticing, as many of my 
textures have a filtered layer with normal on top. At least now I'm 
aware that it happens, and how it affects the artifacts... :)

--
Jaime

Post a reply to this message