Mike the Elder wrote:
> I, cannot, however endorse your admonition to stay away from blinds. Our very
> survival depends on window blinds.  Without them, it would be curtains for
> everyone!  ;-)

This panes me deeply.

This is drape. Just drape.

You shade know better.

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

Who could pollute the entire world?
     -- Nausicaa

Post a reply to this message

William Tracy <wtr### [at] calpolyedu> wrote:

> This panes me deeply.
>
> This is drape. Just drape.
>
> You shade know better.
>
I see you've mustard up some puns that you think
will make you the wiener of this contest.

Although I relish the challenge, I fear you will
never ketchup.

Boast Regards,
Mike C.

Post a reply to this message

"triple_r" <rre### [at] hotmailcom> schreef in bericht 
news:web.47f62829fe9c4eeaae42298f0@news.povray.org...
> "Thomas de Groot" <t.d### [at] internlDOTnet> wrote:
>> we should need to see your radiosity settings...
>
> Sorry.  Of course.  Here are the settings for the high quality render:
>
> radiosity {
>  normal off
>  count 300
>  recursion_limit 2
>  error_bound 0.35
>  pretrace_start 0.08
>  pretrace_end 0.02
> }
>

Like Bill said, you could decrease error_bound to 0.2, but I would increase 
count to 1000 at least. Also pretrace_end should be much lower for better 
results.

I post below the settings I have used for an indoor scene, using 2-pass 
radiosity. Maybe you could try that.

//--- start code ---
    radiosity {
      //saving radiosity data:
      #if (RadSave)
        save_file "DG1.rad"
        pretrace_start 0.1           // start pretrace at this size
        pretrace_end   0.0025           // end pretrace at this size
        always_sample on            // turn sampling in final trace off [on]
        count 1000
        error_bound 0.2
        recursion_limit 2

      //reading radiosity data:
      #else
        load_file "DG1.rad"
        pretrace_start 1
        pretrace_end 1
        always_sample off            // turn sampling in final trace off 
[on]
        count 100
        error_bound 1.8
        recursion_limit 1             // how much interreflections are 
calculated (1..5+) [3]
      #end

      //common settings:
      brightness 1                  // brightness of radiosity effects 
(0..1) [1]
      nearest_count 8               // higher -> higher quality (1..10) [5] 
can go up to 20. little effect on render time higher is better
      adc_bailout 0.005
      low_error_factor 0.8           // reduce error_bound during last 
pretrace step
      minimum_reuse 0.015           // reuse of old radiosity samples 
[0.015]
      gray_threshold 0            // increase for weakening colors (0..1) 
[0]
      normal off                    // take surface normals into account 
[off]
      media off                     // take media into account [off]
      //max_sample 1.0              // maximum brightness of samples
    }
//--- end code ---

good luck! And by the way, you have there a nice image/scene!

Thomas

Post a reply to this message

triple_r wrote:
> I have to confess I cheated.  This is a combination of two renders. 

Nice image! The sad part is that it takes just too much effort to render
it accurately - and still you get errors/can't be sure when the result
is correct. You should ask if fidos could make a reference render of it
with his path tracing Pov.

Post a reply to this message

Well, Triple, I couldn't help you with the artifacts. But I can give you 
another great advise: replace the dog on the couch with a naked, very curvy 
girl laying on the couch, and see full of excitement, how the stripes of 
light play on her curvy back...



"triple_r" <rre### [at] hotmailcom> schrieb im Newsbeitrag 
news:web.47f617807305684bae42298f0@news.povray.org...
>I have to confess I cheated.  This is a combination of two renders.  The 
>low
> quality render has fewer artifacts on the blinds because it just smoothes 
> them
> over, but lots of artifacts around the ceiling.  The higher quality render 
> cuts
> down on artifacts around the ceiling, but causes a lot of artifacts on the
> blinds.  I went all the way up to count=1600 without any improvement. 
> Wherever
> there's a gap there are a lot of artifacts.  Does this mean I have not 
> balanced
> count and error_bound correctly, or does anyone have other ideas on how I 
> might
> improve or fake the rendering quality?  Either way, my tip for the day? 
> Stay
> away from blinds.
>
> - Ricky
>

Post a reply to this message

triple_r wrote:

> I have to confess I cheated.  This is a combination of two renders.  The
> low quality render has fewer artifacts on the blinds because it just
> smoothes them
> over, but lots of artifacts around the ceiling.  The higher quality render
> cuts down on artifacts around the ceiling, but causes a lot of artifacts
> on the
> blinds.  I went all the way up to count=1600 without any improvement. 
> Wherever
> there's a gap there are a lot of artifacts.  Does this mean I have not
> balanced count and error_bound correctly, or does anyone have other ideas
> on how I might
> improve or fake the rendering quality?  Either way, my tip for the day? 
> Stay away from blinds.


Nice :)

Will you share the source perhaps?

There are afair patches to have over 1600 samples, and/or have them better
distributed

Post a reply to this message

=?UTF-8?B?UmFmYcWC?= Maj <no.### [at] invalidcominvalid> wrote:

> Nice :)
>
> Will you share the source perhaps?

I would be glad to, but it's a little big.  In the interest of time I went with
image maps and meshes rather than the purist POV way.  Don't have time at the
moment, but maybe I'll scale down the images, save some lower-resolution meshes
and post it.

 - Ricky

Post a reply to this message

"triple_r" <rre### [at] hotmailcom> wrote:
> I have to confess I cheated.  This is a combination of two renders.  The low
> quality render has fewer artifacts on the blinds because it just smoothes them
> over, but lots of artifacts around the ceiling.  The higher quality render cuts
> down on artifacts around the ceiling, but causes a lot of artifacts on the
> blinds.  I went all the way up to count=1600 without any improvement.  Wherever
> there's a gap there are a lot of artifacts.  Does this mean I have not balanced
> count and error_bound correctly, or does anyone have other ideas on how I might
> improve or fake the rendering quality?  Either way, my tip for the day?  Stay
> away from blinds.
>
>  - Ricky

You might want to consider, if you haven't done it, adding some ambient value to
the drapes and the rooms materials. I've had some success in eliminating those
artifacts without having to ramp up the radiosity quality levels to extremes.

Post a reply to this message

"fls13" <fls### [at] netzeronet> wrote:
>
> You might want to consider, if you haven't done it, adding some ambient value to
> the drapes and the rooms materials. I've had some success in eliminating those
> artifacts without having to ramp up the radiosity quality levels to extremes.

Indeed there is about 0.1 or 0.15 for ambient on the blinds, if I recall.  Adds
some 'translucency' too.  Seems to be one of those scenes where low quality
radiosity is worth the tradeoff of fewer artifacts.  Actually the best luck I
had was with the pseudo-path-tracing method (change reflection to randomly
diffuse reflection).  Worked great with just a few minutes for setup, although
it sure took a while--and no textures, of course.  Thanks for the suggestion.

 - Ricky

Post a reply to this message