> Alain<aze### [at] qwertyorg>  wrote:

>>> I'm working on a concept for a shoreline backed by Basalt columns. I'm using a
>>> randomly tiled hexagonal pattern function with a height field and multiplying it
>>> by f_ridged_mf
>>>
>>> The problem I'm having is that the tops of the columns should be flat, not
>>> slanted, so that the base offset follows the contour of f_ridged_mf, but their
>>> height does not.
>>>
>>> Not sure if that makes sense.
>>>
>>> Any advice would be appreciated.
>>>
>>> Regards,
>>> A.D.B.
>>>
>>
>> For that, I'd use an hexagonal prism, modeled as a mesh, that is tilled
>> with various vertical scales dictated by your function.
>>
>> The mesh is to reduce the memory use and will render much faster than
>> the isosurface version.
>>
>>
>>
>> Alain
>
> Yeah... Not sure what you have in mind, but I couldn't figure out anything that
> looked remotely right.
>
> I realize there are going to be problems with the height_field approach, not the
> least of which are the vertical striations in the columns, I would like to use
> your approach.  (although I might use a csg hex tile)
Using CSG is correct, but you won't benefit from the instantiation you 
get with meshes, resulting in a increased memory useage.
>
> I could create caves this way.
>
> the thing I like about the hf approach is that I could conceivably create a
> pigment function that contains multiple sub-patterns using pigment_map (Basalt
> columns turning into black sand, for instance.)
>
> I'm quite at a loss...  What I'd like to do, ultimately is a shoreline of basalt
> columns and outcroppings fading into green hills inland, which is why I was
> using the height_field.  I tried using an isosurface, but the max_gradient was
> way too high. (over 5000)
>
> Regards,
>
> A.D.B.
>

In real life, those columns don't "fade" into some other features, they 
get covered up with the other features.
So, you can have a hightfield for your sandy area at the base, and 
another for the grassy regions.



Alain

Post a reply to this message