> Stephen <mca### [at] aolcom> wrote:
>> On 21/07/2013 9:37 PM, MichaelJF wrote:
>>> This will take the one or other day due to the rendering times. I have reviewed
>>> the two lower levels now. Not completely satisfying but the best compromiss I
>>> can yield so far. Sometimes a change of 0.01 for the scattering color changes
>>> the image completely having this strange density maps. I'm not quite sure to
>>> repair the density of the upper "cloud", may be I change the color to be more
>>> white. But as Thomas said the lighter parts are blown up into the air and they
>>> must be more easily be influenced by the wind, which is blowing here from right
>>> to left, especially at greater altitudes.
>>
>>
>> We have the patience. :-)
>>
>> --
>> Regards
>>       Stephen
>
> Sometimes I'm not quite sure I will have the patience needed. But today I found
> out the reason of my problems just by chance. I always had ribbon like
> structures around my "cloud" which were annoying completelly. But they did not
> come from the df3-file or the scattering or absorbing parameters or the the warp
> turbulence, I would have expected them. They came from "interpolate 2" within
> the density file definition. I ever thought that a cubic interpolation would be
> better than a linear one only resulting in longer rendering time. But seems, I
> was wrong with that.
>
> Best regards,
> Michael
>

It's a bug present in version 3.6.x. When you have areas with a value of 
zero, the interpolation cause some areas to evaluate to a negative value 
that warps around to give values that are close to 1. I think that it's 
corrected in version 3.7.

The workaround is to never have any zero in the DF3 file and to adjust 
the colour_map accordingly.

Yet another reason to upgrade to version 3.7 :)



Alain

Post a reply to this message