AstralScourge
commented
3 days ago If normal-sized noise is used, you could probably also make the savannas spawn at slightly higher temperatures. Seems like small ones spawn all over regardless of temperature noise size, and I know that you cited that as a problem you were trying to avoid with savannas. I assume that their current spawn frequency was part of an attempt to make sure that they weren't as difficult to find as the other hot biomes, but if hot zones are more common, then that shouldn't be an issue anymore.
Tslat
Cruuk
My knowledge about this is definitely not comprehensive, but I did test it out. Was going to wait a bit to report this, but when I learned that the apparent fix was simple, I figured I might as well do it now
Deserts in Precasia are known to be really sparse (to the point where /locate often fails to find them). When they are found, they're usually really large, often much larger than any vanilla biome would be (including vanilla deserts).
The previous hypothesis around this was that vanilla was the same way, but in Precasia the deserts were just taking the entire "hot zones" up, while vanilla had other biomes to blend in with the deserts, making it less apparent in vanilla.
However, this was inconsistent with my experiences playing vanilla, during which "hot zones" (and significant ones, too, not just an isolated small savanna or something) as a whole and not just deserts were fairly easy to come by, and there was pretty much always a desert within /locate range. If this was really the same as vanilla, then overworld hot zones would be around as hard to find as Precasian deserts (which was definitely not the case).
The main issue with testing this is that AMIDST and similar tools don't exist for Precasia worldgen, but that's alright, because I made my own version, more or less. It basically just loops through the coordinates within a (really large) range and then writes an image pixel by pixel using the biome at the location. The biome at the location is found using the same method that /locate uses to get the biome without having to load the chunk. I also tested in-game to make sure this was accurate
This allowed me to actually test how often the biomes actually occurred instead of relying on /locate anecdotes. Just inserted this code into a command in a WikiHelper dev environment
For Overworld hot zones, I included these biomes:
And for Precasian hot zones, I included these. I know jungles are a bit more temperate, but since their color is so dark, they shouldn't detract from desert + savanna anyway:
The scale is 1 pixel = 32 blocks (was "supposed" to be 8 but I didn't know there was a factor of 4 used for querying these biome noise sources), so the entire image is around 130k blocks on each side. That's too large for GitHub, though, so I had to just take a screenshot of the zoomed out image instead of uploading the real image
Results of this:
Overworld biomes:
Precasian biomes (same seed):
Precasia hot zones (desert + savanna) were indeed significantly larger than Overworld hot zones on average. Indeed, in many cases, Precasian deserts alone would be larger than the entirety of the smaller Overworld hot zones, with only the largest of the Overworld hot zones competing. These were longer and curved more, as opposed to the Precasian ones that just looked like larger clumps. Even these would often still get dwarfed when you added the surrounding savanna in. It is also fairly common to just not have a desert at all within not just 30k or more blocks.
Overworld hot zones were smaller, more common, and less sparse. They followed a similar pattern in which the desert/badlands were generally round clumps in the center, and that they were surrounded by savannas, but that was the only similarity.
Honestly, this seed didn't exemplify the size issue the best, but it works well enough in showing the difference between the hot zones of the two biomes. This should disprove the notion that the difference is solely a result of other overworld biomes creating variance with the deserts. I'd had other seeds that exemplified this better before and I didn't want to keep resetting until I found another such seed. The desert on the right is still much larger than most hot zones, with the only hot zones that can compete doing so because they twist and turn more (while the precasian desert is just a huge clump).
Ultimately, this is still comparing ALL hot biomes in the overworld with just the deserts in precasia, so there should be more of a difference anyway.
I remembered noticing something before about a large biome setting being used for Precasia worldgen. After looking into this, it turns out that Precasia worldgen uses minecraft:temperature_large for its temperature noise. https://github.com/Tslat/Advent-Of-Ascension/blob/abc9126ba2016e2ca9292ecfcca147c202ad891f/data/aoa3/worldgen/noise_settings/precasia.json#L250
Checking the vanilla Minecraft noise settings, the overworld uses minecraft:temperature, whereas large biomes uses minecraft:temperature_large (like precasia).
This would explain why the hot zones are so hard to find in Precasia, but really large when they're found.
Solution: seems to be to just change the noise used to minecraft:temperature (i.e. not _large). (The noise for vegetation is also large? IDK what that does, but I figure it should probably be consistent with overworld normal-sized biomes unless there's a specific reason to have it like that).
Obviously, I can't tell what exactly any of these do just by looking at the code, so I went and tested it myself. I went into the jar, changed these values, and ran the biome finder again. Here's the biome distribution after doing this (same scale as the previous ones):
This looks much healthier than the previous setup. Did some in-game testing and it looked fine there too. In all cases, I found a desert and volcano within the lower half of /locate range. They're not too common, though - there's plenty of spots on the map that are far from any desert (just not ridiculously far).
A worry I had was whether or not this would make tar pits smaller, but this actually doesn't appear to affect tar pit size at all - showing some comparison images here:
Current/large noise:
Normal noise (same seed):
If you look at them closely, while the other biomes move around, the tar pits are in exactly the same location relative to the spawn (marked with magenta). They're also the exact same size and shape. I assume that this means that tar pit placement is independent of temperature, which makes sense given how it can spawn in both deserts and the more temperate biomes.
I'm assuming it's not, but if the current setup happens to be intentional, I'd argue that this is the better option, since all the current settings really do is make it harder to find deserts, which is even more painful in Precasia since there's not even any generated structures along the way. And let's face it, most players aren't going to find a bone horn in an attercopus nest with its current drop rates. This should also make it easier to add future resources specific to hot biomes.