On the new game, I want to do a third person, open world action RPG, with a over the shoulder camera. Think Soulborne games or third person Elder Scrolls games.
Anyways, this new viewing perspective and player freedom means that the environment is far more visible and important. As such, I like to have more diverse and interesting terrain. At the same time, I need some method to quickly create those said terrain; AND give them the right collision behavior; AND make them navigable by the AI.
The first thing that comes to mind is voxel terrain, or Minecraft. If everything is in a grid, then it simplifies a lot of the math underneath. So let's start with that. With some amount of effort, this is the result:
(PLACEHOLDER ART)
Not too bad huh? This picture doesn't shown it but the engine already replicated most of the Minecraft features, such as "infinite" terrain extending towards all directions, creating/removing blocks, and even Minecraft's unique "lighting" system.
But of course, there is a pretty big problem: Cubes. Lots of cubes. Lots of not exactly pleasant to look at cubes. So the next step is attempting to smooth out the cubes:
After considerable amount of effort, this is the result. Looks fairly solid at first glance. But there is a severe artifact on diagonal "cliffs":
Unfortunately, after even a greater amount of effort trying to fix it, I concluded that this is a limitation due to the constraints of the grid system. I am not sure how to solve it, or that it is even possible to fix.
Much smoother.
Unfortunately, height maps are only 2D, and they don't allow things like caves, or tunnels, or buildings... So this is where I am stuck now.
Something I considered is use the heightmap for the base terrain and the voxel grid system for things that are build on top of the terrain. It will take some effort to integrate the two separate system together tho. Or perhaps just use the smooth voxel system, accept it's limitations and call it a day? Or something else? Not sure yet. Currently weighting the options...
What about Heroine Rumble's public version? We've been waiting for it for quite a long time, and I'm starting to feel like you're forgetting about us
ReplyDeleteIn my opinion, I would use the voxel system and call it a day. Unless you can line up additional assistance (either via volunteers or by getting immensely popular and partnering with a company such as tinybuild), it probably isn't worth it to sweat every last detail. I know I can't speak for everyone, but I didn't come to this genre of games to stare at terrain. I won't notice if there are artifacts on hill or cliffsides, as I'll be otherwise entertained. Stick to what you're good at and don't worry endlessly about what you're not.
ReplyDeletePS. While I wholeheartedly support the idea of getting a new game underway now that you have some experience under your belt, the public version of HR is lagging a full two builds behind the patreon version. If you want to move to a paid-only model, that's perfectly acceptable and respectable, but you should inform your audience as such. If, as I suspect, it is entirely due to sussing out bugs and enthusiasm for the new game, then no harm no foul. Again, just make sure to keep all sides informed. Have a great week:)
The diagonal hills don't look so bad, heck I wouldn't notice if you haven't pointed it out. Besides, currently you are just showing us a plain terrain, once it will have other stuff (rocks, trees, grass and etc), it won't be so noticeable
ReplyDeleteThis is a large undertaking, and I wish you the best of luck!
ReplyDeleteFrom what I can see, if you truly spend time on this upcoming game it will be rather amazing.
But it will also take a decent amount of time to complete.
Good luck, and stay safe!