[Question] What's the range of the example cave generation nodes tree?

[Filocava99]

Hi, I was trying out your tree example for cave generations (base64: EwCamZk+GgABEQACAAAAAADgQBAAAACIQR8AFgABAAAACwADAAAAAgAAAAMAAAAEAAAAAAAAAD8BFAD//wAAAAAAAD8AAAAAPwAAAAA/AAAAAD8BFwAAAIC/AACAPz0KF0BSuB5AEwAAAKBABgAAj8J1PACamZk+AAAAAAAA4XoUPw==) but I'm a bit confused on the values range and when it is supposed to be hollow. Is it solid on the min/max values or when it's very close to 0?

[Auburn]

In the NoiseTool the default setting is any value >= 0 is solid

You can see the generated min/max values and change the isosurface value in the debug window

[Filocava99]

Mmm that's weird cause I am getting a very weird result (https://prnt.sc/XPb80X4USDDK). The max value that I get from the NoiseTool is a 0.02, almost all the values are negative.
For the sampling I am using fnGenSingle3D and iterating each block of a cube (16x16x256) , it's not efficient but I'm just doing some tests.
Is there any detailed documentation for the parameters of the Gen* functions (grid, single, etc)?

FastNoise2.INSTANCE.fnGenSingle3D(caveNoise, chunkX*16+x.toFloat(), y.toFloat(), chunkZ*16+z.toFloat(), seed)
Just for reference, I made a binding using JNA, it works fine but maybe I understood the parameters wrong.

[Auburn]

Sorry I got that the wrong way round, value >= 0 is air.

Max is only 0.02 to keep the cave sizes smaller, you can change this by adjusting the Subtract node.

[Filocava99]

Hi, thanks for replying so quickly.
I tried that but the result it's the inverse of what you see in the picture (the holes are now filled and viceversa).
At this point I am wondering if I am using the generator functions correctly.

Given a cube X * Y * Z of size 1625616, running the following code

for(int x = 0; x < 16; x++){
  for(int y = 0; y < 256; y++) {
    for(int z = 0; z < 16'; z++) {
      bool isCave = fnGenSingle3D(caveNoise, x, y, z, seed) >= 0f
      ... store result ...
    }
  }
}

it should give me a coherent list of coordinates that form up a cave.

I also tried a simple terrain generation tree and instead of sampling each 2D coordinate I tried using GenUniformGrid2D; I tried generating two chunks, which means I run the following code:

int chunkX = 0, chunkZ = 0;
float[] chunk1 = fnGenUniformGrid3D(chunkNoise, arrayPointer, chunkX*16, 0, chunkZ*16,16, 256, 16, 0.1f, seed, null)
int chunkX = 1, chunkZ = 0;
float[] chunk2 = fnGenUniformGrid3D(chunkNoise, arrayPointer, chunkX*16, 0, chunkZ*16,16, 256, 16, 0.1f, seed, null)

In that case the surface of the terrain is coherent (meaning it doesn't make any weird jump) only inside the same chunk, making highly noticeble the connection between the two chunks.
When using UniformGrid how do I specify the steps it makes with the coordinates? If x is 0 and xSize is 16, what does it use a step for reaching the required size? 1? Or does it use the frequency parameter?

I also wanted to describe the issue with the terrain generation cause everything is problably related to how I am using the APIs given I couldn't find any documentation.

[Auburn]

It's likely caused by you indexing the noise array using the incorrect axis, in your first code block above your for loops go x, y, z when it should be z, y, x. Have a look at the example code on the wiki.

As for the cave noise, I'd guess the issue is somewhere else in your code, what you have posted looks correct.

[Filocava99]

Got the cave noise to work using GenUniformGrid instead of GenSingle3D. Weird tough