Christian Froeschlin <chr### [at] chrfrde> wrote:

> there is no special treatement for image_maps. If the value
> you put before the image_map entry is V, and the boxed pattern
> at a certain point in space evaluates to x (0<=x<=1), then the
> resulting pigment is:
>
>   x ==  0  ==> Fully transparent
>   x >=  V  ==> Pigment as specified at V (last entry in map)
> 0 < x < V ==> Resulting pigment is weighted average of
>                fully tansparent and specified pigment
>                depending on how close x is to 0 or V.

All true, of course. Basically like 'spherical.' BTW, what I have a somewhat
difficult time with (always, it seems) is the 'order' of the index list for
these kinds of patterns. Just a difficulty in keeping the spatial ordering in my
head.
[0.0] // *outside*
[1.0] // *inside*
My own way of visualizing this is...backwards. I.e., 0.0 tells me its the
origin, and 1.0 looks like 1-unit from the origin and everything outside. I have
to *re-think* every time I use such a pattern. I need to cement into my brain
that the values really mean 'weight' or intensity, leading away from the
origin--the origin having a weight of 1.0  (That may not even be the correct
paradigm--but it helps me.) What screws me up is that gradient y (or x or z) has
an index list that implies [0.0...] *as* the origin--the pattern goes up from
there.  And that's the first pattern I learned to use. Of course, the docs
describe each of these patterns correctly--but I still have a hard time.

>
> E.g. if x = 0.5*V you will get 50% transparency. For your
> decal you don't want fading from the center so you need a very
> small value for V.

That's the little bit that threw me off initially.
>
> > Note the horizontal wing, and how the image is barely visible on
> > it--I scaled the box very thin to get that.
>
> if possible, it will of course be preferrable to apply the
> texture to the target object *before* adding extra geometry,
> so it will not be visible at all on the horizontal part.
> But I assume this was just for demonstration ;)

Yes, just as a demo. I completely agree: If possible, such a decal should be on
only one object, before being made part of a larger CGS construct. My scene is
really meant to show what is possible even in the worst cases. And it was fun to
do.  ;-)

My B-29 bomber model isn't textured 'correctly' (neither is this simple example
scene)--the major texture in both cases is just a regular 3-D cells pattern,
applied once to the entire model, rather than its parts. (And it shouldn't be
3-D in any case, but rather wrapped around the parts using uv-mapping or some
such.)  But that being the case, this decal idea came in very handy, as a simple
addition to the entire object.

Ken

Post a reply to this message