Le 16-08-07 à 11:28, Stephen a écrit :
> On 8/7/2016 4:13 PM, Stephen wrote:
>> I would post the code but
>
> I had better mention that I need to use a multiplier for each media to
> balance the colours.
> E.g.
> //------- Box0 Raw Script Begin -------
> #declare Emission_Factor_R = 10.0 ;
> #declare Emission_Factor_G = 3.0 ;
> #declare Emission_Factor_B = 15.0 ;
>
> #declare df3_Scarab_ =
> material{
>   texture {
>     pigment {
>       color rgbft <1.000,0.000,0.000,0.000,1.000>
>     }
>   }
>   interior{
>     ior                 1.000
>     caustics            0.000
>     dispersion          1.000
>     dispersion_samples  7.000
>     fade_power          0.000
>     fade_distance       0.000
>     fade_color          rgb <0.000,0.000,0.000>
You shopuld remove the 7 lines above. They are useless and may introduce 
some problems.
>     media {
>       method     3
>       intervals  20

Using intervals larger than 1 HUGELY increase the rendering time and can 
cause artefacts.
NEVER use anything else that intervals 1 with method 3.

>       samples    3//, 6

Instead of increasing intervals, increase samples. samples 100 is *much* 
faster that intervals 20 samples 3 (for 60 total samples) for a finer 
result.
The ", 6 " is totally useless and silently ignored.

>       confidence 0.900
>       variance   0.008

confidence and variance are used ONLY with method 1 and 2. With method 
3, they are meaningless and better left out.

>       absorption rgb <0.000,0.000,0.000>
>       emission   rgb <1.000,0.000,0.000>  * Emission_Factor_R
>       aa_threshold 0.100 // default value
>       aa_level    4 //default value = not needed
>       density {
>         density_file df3 "F:\Graphics\B3D
> Data\Scenes\Dahlgren\Eastern\B\Shenlong_01A_R_.df3"
>         interpolate 2
>       }
>
>     }
>
>     media {
>       method     3
>      // intervals  20
>       samples    80
>      // confidence 0.900
>     //  variance   0.008
>      // ratio      0.900

ratio is also useless, and ignored, with method 3.

>      absorption rgb <0.000,0.000,0.000>
>          emission   rgb <0.000,1.000,0.000> * Emission_Factor_G
>       aa_threshold 0.100
>       aa_level    4
>       density {
>         density_file df3 "F:\Graphics\B3D
> Data\Scenes\Dahlgren\Eastern\B\Shenlong_01A_G_.df3"
>         interpolate 2
>       }
>
>     }
>
>     media {
>       method     3
>
>       samples    80
>
>
>
>       absorption rgb <0.000,0.000,0.000>
>       emission   rgb <0.000,0.000,1.000> * Emission_Factor_B
>       aa_threshold 0.100
>       aa_level    4
>       density {
>         density_file df3 "F:\Graphics\B3D
> Data\Scenes\Dahlgren\Eastern\B\Shenlong_01A_B_.df3"
>         interpolate 2
>       }
>
>     }
>
>   }
>
>   translate <-0.500,-0.500,-0.500>
> }
>
> //------- Box0 Raw Script End ---------
>
>
>
>

In method 1 and 2, confidence, variance and ratio are used to control 
how many samples are to be used in the different intervals depending on 
the density and illumination in the various intervals. With method 3, 
using more that a single interval is highly counterproductive.

I've done my tests with intervals and the negative impact they have on 
the performance is astounding.
I demonstrated that samples 100 is way faster than intervals 10 samples 
1 for greatly better results.



Alain

Post a reply to this message