0
\$\begingroup\$

I am making a game, and I want to have life-like terrain generation. Some issues with procedural noise, and other algorithms like midpoint displacement is that some features of the land that they create are inaccurate. For example, continents tend to form with them having large mountains and flat land inbetween, which noise cannot do well.

Are there any existing algorithms that can do something similar to what I am looking for?

\$\endgroup\$
2
  • 6
    \$\begingroup\$ How lifelike exactly do you want it to be? If you want to convince a geologist, then you might need to simulate plate tectonics and erosion. Or do you just want more diverse geography for a more varied game experience? In that case, if noise isn't the solution, then you aren't using enough of it. \$\endgroup\$
    – Philipp
    Commented Apr 2 at 7:41
  • 4
    \$\begingroup\$ I'd agree with Philipp that we need more parameters here. There's too much variation in real world terrain for us to fit instructions to generate all of it to perfect fidelity in a StackExchange answer. But if you just need help with one feature at a time like how to mix different biomes such as ocean, mountain, and plains, we have some past Q&A on that, which may help you. \$\endgroup\$
    – DMGregory
    Commented Apr 2 at 10:01

1 Answer 1

-1
\$\begingroup\$

Games like Minecraft use the Perlin noise or Simplex noise, but it depends, as said in the comments, on how realistic you want the terrain to be. Here you can find more about the Perlin noise usage. Here you can find more about the Simplex noise. Those noises are used to fix parameters like height, temperature... The higher the value in the set point, the higher the terrain will be.

\$\endgroup\$
7
  • 3
    \$\begingroup\$ Perlin noise is obsolete for most applications. It should be replaced by Simplex Noise, which fixes most of the problems with Perlin Noise (like directional artifacts) and is faster to calculate as well. It was developed by Ken Perlin himself as a direct upgrade to his older algorithm. Besides that, I think this is a poor answer because it doesn't elaborate on how exactly noise should be used to achieve what the question author wants. They already said that they are using noise, so this answer is probably of little value to them. \$\endgroup\$
    – Philipp
    Commented Apr 2 at 8:46
  • \$\begingroup\$ @Philipp I gave a link to how to use the noise. \$\endgroup\$
    – Qualcuno2
    Commented Apr 2 at 8:55
  • 1
    \$\begingroup\$ None of these links describe how to solve OP"s problem in trying to get convincing continent generation from a noise function. \$\endgroup\$
    – DMGregory
    Commented Apr 2 at 9:58
  • 2
    \$\begingroup\$ Those noise types, on their own, do not generate realistic terrain, and you have not provided an algorithm that does so using them. The hard part is that algorithm: how to sample and layer the noise to mimic different terrain types. \$\endgroup\$
    – DMGregory
    Commented Apr 2 at 10:08
  • 5
    \$\begingroup\$ Personally, I would not upvote an answer that just links to an existing answer without adding any clarity or guidance of its own. So if creating high-quality answers is your goal, I think you can do better. Imagine you yourself were struggling with noise outputs that look like unrealistic endless mountain ranges: what advice would you hope another user would share with you? Would you want links to pages you might have already read, or a detailed explanation about how to fix your specific problem? \$\endgroup\$
    – DMGregory
    Commented Apr 2 at 10:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .