- Home /
Perlin Noise discrepancies
I am using Perlin Noise to create a terrain, with great success.
I can effectively "zoom-in" on my terrain by changing the noise scale, and I can move the terrain on a 2D plane by changing a Vector2 for the sample position.
This almost works exactly how I want. However, even though the general shape of the terrain remains the same, there are obvious areas where the noise is not holding together properly.
For instance, my code can increase the number of terrain chunks, effectively increasing the size of the terrain. When I do this, the same issue occurs, and certain peak-regions are getting "smashed" down to lower values. This does not make sense, as the samples are coming from the exact same noise-map positions. Also, there are different parts of the map, where doing these scalings does NOT seem to create the issue? It has really got me stumped.
Not sure if this is just some kind of drawback of Perlin Noise? I am also looking into some kind of fractal-math solution to achieve the affect I desire, in the event no one has any good ideas.