Hi, I’ve been away for a while, but lately I’ve been tinkering with WM again, and I’ve thought of a feature that might come in handy for those who, like me, like to work with games’ worldspaces.
So, this is my idea.
Many game engines have serious problems with steep slopes, especially when they’re over 20 meters tall, nearly vertical leaps; furthermore, some of them also have mid-low poly terrain resolutions, which makes very detailed maps nearly unusable because of the resulting vertex spikes.
What I used to do in the past, when making worldspaces for Oblivion and Skyrim, was rendering the terrains at a low resolutions (2048 or 4096), and then use them as file inputs to re-render the terrain at the actual resolution I needed, which was (16k x 16k).
This would eventually soften all the features, especially when coupling the file input with a blur filter or a similar effect.
Would it be possible to make an output/scalar object capable of scaling up/down the output resolution?
Like, the project’s base resolution might be 2048x2048, while the output would result in a 16384x16384; all the objects until the final output would process the maps at 2048, but the output would re-process the final map at 16384.
ACTUALLY, it would work better as some kind of “inter-objects re-sizer”, like an image processor that takes an object’s output and gives out a resized/re-rendered version of it.
Something like:
[Base resolution: 2048] -> [Advanced Perlin] -> [Erosion] -> [Resizer: 16384] -> [File Output: RAW map]
This way, one would save lots of RAM and time; and this is just a very simple example. One could calculate complex maps at low resolutions and then resize them to, like, filter other objects’ higher resolution outputs (when super details are not needed).
It would work like a crossbreed of the [Height Output] and [File Input] objects, but without an actual disk output/input (not needed, since the map would still get loaded onto the RAM anyway).
Silly idea? Someone already implemented this in a plugin/macro?
PS: an alternative would be giving every existing object a built-in [Output Resolution] control, but it’d be a nightmare to program, I guess.