|
 |
Le_Forgeron <lef### [at] free fr> wrote:
> <0,0,0> is the only 100% BAD value.
>
> Remember inside_vector is not a position, it's a direction.
After I posted my hasty response, I realized that even if it WAS a location,
it's a torus, and that would be bad too :D
<1, 0, 0>*(MajorRadius) would be a good vector that would intersect the torus if
the test ray originates at the origin.
I would presume that if that inside_vector is part of the mesh definition, then
when the torus gets rotated, the inside_vector gets modified in the same
fashion.
It's always these little details. ;)
Post a reply to this message
|
|
|
|
Am 05.08.2016 um 19:03 schrieb Bald Eagle:
> <1, 0, 0>*(MajorRadius) would be a good vector that would intersect the torus if
> the test ray originates at the origin.
The `inside_vector` has no relation to /any/ points in space whatsoever;
it is a mere direction vector. Even the length does not matter.
> I would presume that if that inside_vector is part of the mesh definition, then
> when the torus gets rotated, the inside_vector gets modified in the same
> fashion.
Effectively, yes.
Technically, no -- rather than actually modifying the entire mesh data,
rotations and other transformations are instead just accumulated in a
pair of matrices, and whenever POV-Ray needs to do math on the mesh it
uses these two matrices to convert input and output data of the
algorithm, so that it can do the math on the original untransformed mesh.
Post a reply to this message
|
|
