+- Cuberite Forum (https://forum.cuberite.org)
+-- Forum: Cuberite (https://forum.cuberite.org/forum-4.html)
+--- Forum: Development (https://forum.cuberite.org/forum-13.html)
+--- Thread: Some semi-random thoughts on terrain generation (/thread-409.html)
RE: Some semi-random thoughts on terrain generation - NiLSPACE - 11-14-2014
As far as I understand the Bresenham algorithm (I could of course be wrong) can only make straight lines, but the way I'm doing it I can make the branches go in a direction more than another one.
Code:
X X---
----X--
------X
------XRE: Some semi-random thoughts on terrain generation - xoft - 11-14-2014
Oh, right. And do you use it at all?
RE: Some semi-random thoughts on terrain generation - NiLSPACE - 11-14-2014
Well.. yes the branches sometimes go down.
RE: Some semi-random thoughts on terrain generation - xoft - 11-17-2014
I've just spent a day implementing the Improved Perlin noise ( http://mrl.nyu.edu/~perlin/noise/ ) in the hopes that it would generate the noise faster than our current cubic generator. Pff. It's more than 10 times slower! It seems that I've optimized the Cubic noise down to its bones. I'm not sure if I want to spend another day optimizing the Improved noise only to find out it makes no difference
RE: Some semi-random thoughts on terrain generation - xoft - 11-17-2014
Nope, it won't get any faster than cCubicNoise.
RE: Some semi-random thoughts on terrain generation - NiLSPACE - 11-17-2014
Aww.
RE: Some semi-random thoughts on terrain generation - xoft - 11-18-2014
Hah! It's for real, going for a walk on fresh air lets you think more clearly
I've been using the Improved Perlin noise wrong, they're each good at something slightly different, and we may yet find some improvements. However, the noise generated will be different, and I don't want to break existing generators, so it'll need to go to separate classes.
RE: Some semi-random thoughts on terrain generation - xoft - 11-19-2014
Nice going so far. I've implemented some ideas from the ImprovedNoise into a 3D noise generator (InterpolNoise) and it's now as fast as the CubicNoise 3D generator. *BUT* CubicNoise is highly optimized, InterpolNoise is not yet optimized, and it should be possible to optimize it pretty well.
RE: Some semi-random thoughts on terrain generation - xoft - 11-19-2014
And the numbers are in. For the (non-composable) Noise3D generator the InterpolNoise is about 4 times faster than CubicNoise. This means we could increase the number of octaves, make the noise look better and still have a faster generator. I've pushed the changes to the ImprovedNoise branch, go check it out if you want.
I have a feeling, too, that the new noise has a shape slightly better for generating nice overhang shapes in the composable Noise3D and BiomalNoise3D generators. We'll see when I change the noises used in those. Tomorrow.
RE: Some semi-random thoughts on terrain generation - NiLSPACE - 11-19-2014
I'm really curious how it will look
