Even 3D Plugins such as agoaj's one here that demand WebGL require a lot of code to be written for handling the 3D mathematics and environment, the amount of work required is huge. Canvas2D does not have 3D support, so to code 3D with this requires a heck of a lot of code, most developers will start going down this rabbit hole then quickly regret it. However, MV also runs on platforms that do not have WebGL support, so Canvas2D (software graphics, no graphics card used) needs to be supported too. MV can handle "mode 7" (and 3D in general) far better than previous vanilla RPG Maker versions - this is thanks to the native support for WebGL. I've seen some solutions that do this, but for the end-user it's extra bloat and it locks out some options (Canvas rendering one of them, so they're definitely not an option for me).ĮDIT: You've got me super excited about all this again! Thank you!Ĭlick to expand.The story is a little more complicated than this. I agree about avoiding 3D Javascript libraries. I am very interested in how you tackled these issues (if you tackled them at all, in some cases it's just faster to draw the whole thing without worrying, especially when it comes to Javascript performance). I'm still thinking about whether I want to generate meshes of repeating tile textures at run-time, which would mean exploding the tileset into a 3D texture at load-time to take advantage of hardware texture repeat and hardware tile animations. A clever frustum culling system would probably be best (quad tree should work super nice). Javascript isn't very good with 3D mathematics. I'm not sure what you mean by "vertex limit" - are each of your tiles an individually uploaded mesh? I still need to think about how I'm going to do the tile rendering. MV is definitely optimised for rendering to a 2D surface that can be panned without updating, not a free-moving 3D surface, so a lot of recalculations happen each time the player takes a step. I don't think I'd use MV's tile-mapper for updating the tiles or calculating visibility, I'd roll with a completely custom approach. I am really excited to see your Plugin, really can't wait to see more. Will you also be exploring the possibility of having 3D models on the map? This is something on my feature list. I'm interested if you'll be sticking to the 3D grid movement or if you will also implement a new movement system?
![rpg maker mv 3d labyrinth rpg maker mv 3d labyrinth](https://i.ytimg.com/vi/DTj6BsUvx5U/maxresdefault.jpg)
When I was working on my 3D maps Plugin I realised that there's a lot of components to be built to make a really flexible Plugin, the main one being the character movement, as such I'm working on a character movement Plugin before moving onto 3D maps.
![rpg maker mv 3d labyrinth rpg maker mv 3d labyrinth](https://cdn.akamai.steamstatic.com/steam/apps/1957960/ss_41d5ef05667dcdb4327468541e86eae7896fb41c.1920x1080.jpg)
I'm guessing right now you're taking the rendered tilemap and using it as a texture on a free-moving 3D mesh? This is what I explored in my early attempts, but overall I think it may be better to do a batched tile renderer, the logic being that if you're drawing the tiles to the tilemap, you might as well draw them to the screen in 3D and avoid the middle-step. I am particularly impressed with the mouse-look.Īs for the WebGL requirement, a while ago I made some experiments with rendering 3D models in canvas: that might give you some ideas on making a compatibility mode to run on instances without WebGL support (a lot of Android devices are in this category still). Hey there, I want to say good job on these preliminary tests - MV is still missing 3D maps despite the engine having native support for hardware acceleration.