andyz
|
The z coordinate in get_perlin_noise does not work as expected for me - its scale seems to be way off that of the x & y axis.
Normally with 3d perlin noise I can swap the x & z coordinates and get rougly similar results but in FF lua get_perlin_noise(x,y,0,10) is totally different to get_perlin_noise(0,y,x,10) because everything is stretched horizontally with the latter.
Why is the z scale different?
|
Posted: August 25, 2014 9:33 am |
Details
E-Mail
|
Indigo Ray
Adam

|
The 4th parameter of get_perlin_noise controls the "scale" of the x and y coordinates, but it does not seem to affect the z coordinate. The noise(0,y,x,scale) is not stretched when the scale is set to 500 or so.
|
Posted: August 25, 2014 8:51 pm |
Details
E-Mail
|
GMM
Moderator
Filter Forge, Inc
Posts: 3491
|
Andyz, please attach or email a filter that would illustrate your issue.
|
Posted: August 27, 2014 11:17 am |
Details
E-Mail
|
andyz
|
here is simple example noiseproblem.ffxml
|
Posted: August 28, 2014 6:02 pm |
Details
E-Mail
|
Vladimir Golovin
Administrator
|
Andyz, yes, the Z coordinate of our perlin noise behaves differently. Our implementation of perlin noise was not intended to provide a full 3D noise. In our 2D adaptation of the Perlin Noise algorithm, the X and Y coordinates are special because the noise must be capable of seamlessly tiling itself along the XY plane, but not in the third dimension. The stretching you see is probably caused by the fact that the Z coordinate is not affected by the scale factor argument.
|
Posted: September 15, 2014 10:28 am |
Details
E-Mail
|
andyz
|
Ok, I am doing something different from the norm (3d noise on a sphere for planets) but I used a LUA implementation of improved perlin noise which is actually very fast!
|
Posted: September 17, 2014 5:34 am |
Details
E-Mail
|