This is no big deal.

I'm experimenting with radiosity (still learning it) and come up with this
image.

No lights, just a "sky sphere" with ambient 1 and the radiosity settings of
the "high quality" example from the advanced tutorial.

Render time 4 hours 34 minutes  50.0 seconds (16490 seconds) on my PIII 500,
196 Mb

I've rendered it with the default AA settings (0.3) but I'm not satisfied
with the result.

In some places the texture lines are jagged or incomplete, and the left side
of the sphere is a mess.

Can it be from the radiosity?

I would like some advice on how to correct those problems, because as these
are my first trys at radiosity Im not quite shure what to try next.

Sorry for the size of the file but the smaller ones have a REAL lousy
quality.


Best regards.

Post a reply to this message

Attachments:
Download 'RadTest.jpg' (149 KB)

Preview of image 'RadTest.jpg'

> I would like some advice on how to correct those problems, because as
these
> are my first trys at radiosity Im not quite shure what to try next.

Radiosity is a tricky beast that you never really master, you always find
yourself tweaking it for the particular scene you are on and you can't
expect the same results twice on different scenes. I'm not quite sure what
you are testing here but the textures you chose are not the best for showing
off radiosity. Most radiosity tests feature scenes with light/pastel or dull
mat textures. Reflection can hide some of the effect of radiosity, though I
do notice it in your scene a little.

Post a reply to this message

> I've rendered it with the default AA settings (0.3) but I'm not satisfied
> with the result.
>
> In some places the texture lines are jagged or incomplete, and the left
side
> of the sphere is a mess.
>
> Can it be from the radiosity?


the best you can do is set AA threshold to 0.0 and you'll see this is almost
a magic zero. Focal_Blur is the alternative but won't be faster. Either the
left side of the sphere needs to be out of focus, or your blur needs
variance 0.0 which is similar to AA 0.0

Rendering in double resolution or so, won't help... My suggestion will be:

+AM2 +A0.0 +R2 -J

Regards,
Hugo

Post a reply to this message

JPGargoyle wrote:
>
> In some places the texture lines are jagged or incomplete, and the left side
> of the sphere is a mess.

Copy & paste from newusers:

  This was rendered with +a0.0 +r9. http://luxlab.com/tmp/aa.png

  Rendered at 500% resolution +a0.3 +r3, Gaussian blur 2.5 pixels in
  Photoshop and resampled (nearest-neighbour) to final size.
  http://luxlab.com/tmp/aa_hq.png

  The most important part is the Gaussian blur. I used 500% resolution
  just to get those very thin lines. For normal a scene 200% to 300% and
  1 to 1.5 pixel blur radius should be enough. The 50% blur radius seemed
  to be the minimum required to prevent visible aliasing. The image will
  look soft. Sharpening filter will just bring back the the jagged edges.
  It's a delicate balance between artefact free and sharp image.

I feel the Gaussian filter produces optimal anti-aliasing. It's so good
it's banned from IRTC :D The example image has maximum contrast. When
local contrast in image is lower less filtering is needed. The default
POV-Ray aa is enough for low local contrast.


This scene and camera position are less than optimal for radiosity
demonstration :) Try lowering the camera near a corner. The contrast
ratio in this scene is naturally low. If you don't wan't to post-process
drop assumed_gamma or use higher value as a kludge to increase contrast.

 
_____________
Kari Kivisalo

Post a reply to this message

"Thomas Lake" <tla### [at] REMOVE-THISshawca> wrote in message
news:3cd9eb9d@news.povray.org...
> .... I'm not quite sure what
> you are testing here but the textures you chose are not the best for
showing
> off radiosity. Most radiosity tests feature scenes with light/pastel or
dull
> mat textures. Reflection can hide some of the effect of radiosity, though
I
> do notice it in your scene a little.
>
>

Well, I was just trying radiosity to see if I could achieve a more realistic
look in that scene, and to learn too :))

Thnx for your comment.

Best regards.

Post a reply to this message

I've tried the setting you suggest and they worked just fine :)))

Thank you very much.

I've never tried different settings for AA (never needed, cause I've always
rendered simple test scenes :P) so your help was very welcome.

I enclose in the next post the resulting pic (without reflections, cause
they took LOTSA resources), just to show the difference.


> the best you can do is set AA threshold to 0.0 and you'll see this is
almost
> a magic zero. Focal_Blur is the alternative but won't be faster. Either
the
> left side of the sphere needs to be out of focus, or your blur needs
> variance 0.0 which is similar to AA 0.0
>
> Rendering in double resolution or so, won't help... My suggestion will be:
>
> +AM2 +A0.0 +R2 -J
>
> Regards,
> Hugo
>
>

Post a reply to this message

Here is the pic.
No reflections this time :))

Post a reply to this message

Attachments:
Download 'RadTest2.jpg' (138 KB)

Preview of image 'RadTest2.jpg'

"Kari Kivisalo" <pro### [at] luxlabcom> wrote in message
news:3CDA721E.A3091134@luxlab.com...
>   This was rendered with +a0.0 +r9. http://luxlab.com/tmp/aa.png
>
>   Rendered at 500% resolution +a0.3 +r3, Gaussian blur 2.5 pixels in
>   Photoshop and resampled (nearest-neighbour) to final size.
>   http://luxlab.com/tmp/aa_hq.png

I saw the pics and I am impressed. This method really achieve great results.


>   The most important part is the Gaussian blur. I used 500% resolution
>   just to get those very thin lines. For normal a scene 200% to 300% and
>   1 to 1.5 pixel blur radius should be enough.

But if I would render this again at 500% or even 200% resolution would take
an ETERNITY for such a simple scene (with all those reflections) and I would
abort the render for sure, cause I want the renders to be as fast as they
can so I can try more things he he he :))

> This scene and camera position are less than optimal for radiosity
> demonstration :) Try lowering the camera near a corner.

I know, but I was just testing, and trying to get a more "real" look, not a
tipical radiosity scene :)

Best regards

Post a reply to this message