> First, thank-you all for your suggestions, they've been most helpful.  In fact,
> I've used almost all of them in one way or another.
>
> It's becoming obvious, however, that I'm going to need the big guns for this
> project, as my last render took... 16 hours..?  My tiny little core duo tablet
> just isn't up to the task of giving me timely results on test renders, even
> without radiosity.
>
> I adapted the method Thomas suggested, using the following settings:
>
> media {
>       intervals 10
>       scattering {
>            5,<0.05, 0.17, 0.3>*0.05
>            extinction 1.0
>            eccentricity -0.23
>            }
>       samples 50
>       absorption rgb<0.95, 0.55, 0.3>*0.05
>       density {
>            function {exp( -(sqrt(x*x+y*y+z*z)-63870)/4 )}
>            density_map {
>                 [0 rgb 0.0]
>                 [1 rgb 0.1]
>                 }
>            }
>       }
>
> Radiosity (The good settings):
>
> radiosity {
>       pretrace_start 0.01
>       pretrace_end   0.005
>       count 600
>
>       nearest_count 12
>       error_bound 0.1
>       recursion_limit 3
>       low_error_factor 0.2
>       gray_threshold 0.2
>       minimum_reuse 0.015
>       brightness 0.58
>
>       adc_bailout 0.01/2
>       }
>
> I know that 1 interval is supposed to be sufficient, and it may have to do with
> my scale, but with only 1 interval I get a ring-effect in the media radiating
> outward from the point of greatest light intensity.
>
> The absorption was added because the original media gave the planet a reddish
> cast.
>
> These settings seem to work fairly well, except that the atmosphere is a too
> dense.
>
> The big guns:
>
> Asus M4A79T Deluxe motherboard
> AMD 3.0Ghz Hexacore processor
> 8Gb ram
> 1 TB Western Digital Caviar Black HD
> Radeon HD 5870 graphics card
>
> Regards,
> A.D.B.
>
>

You must keep intervals at 1.
Instead, you should increase the samples count as needed.
In your case, samples 500 would be significantly faster than intervals 
10 samples 50

The media computation can easily be about 3 times, maybe more, faster 
that way. You can reasonably expect render times under 6 to 7 hours.

Using samples 300 may be enough.

The video card have NO effect at all.

As you scatter more in the blue than other colours, it's normal that 
there is a redish cast similar to what appens for real. Why do you think 
the sun turns bright red at dusk and sun set?
Why does the moon turns brick red durins a lunar eclipse?


Alain

Post a reply to this message