Hey there I am planning to experiment a little with coding and unity’s editor to learn more about Unity’s game dev in general, I am planning to make a hibrid beat em up , using random sprites , and this bet em up level… (the level is not only horizontal, that is interesting! and I like for that reason and here is the motive).
I am wondering 1 -if I need to use a nav-mesh agent to define the playable area? even for player and the enemies? How?
2- Or should I use a raycast system? with different colliders? in the player’s component (many colliders to detect things like ground ,walls, floor ,slopes)? .
3- the level has some sort of slope (where I am shoing in the snapshot.) should I compositive collider?
I am using “compositive collider” and “poligon collider 2D” for now… cause it is simple for me … but I would like to hear comments and experience with this type of game with “slopes”.
If you did not play it you probably do not understand anything.
You can set the environment as 3d, and block impassable areas with invisible walls. This is probably the fastest.
The curious thing about retro beat em ups is that while characters might be technically moving in 3d (because they can jump), the ground may be perfectly flat. Meaning the map may display a slope, but the slope is “painted on the floor”.
To use 3d movement simply set up correct orthographic camera for the scene.
so you suggest to use 3D template? Ok , but how can I deal with this part?
I need to figure out what type of colliders should I use? to make the limits (loyal to the game).
I decided to use Poligon collider, but it is odd and bad for “jump” and it is bad… I need to resolve and planification.
Actual retro games do not ever have precise collision.
In this case I’d likely use 3d world with orthographic camera, stuff some invisible boxes into the area to make it unwalkable, and if it doesn’t work, i’d use mesh collider, likely with conjunction with some CSG toolkit for unity. RealtimeCSG, SabreCSG, etc.
do you suggest me to crop the map? I ve downloaded from “sprite-resources.com” so the level is a mass of piece.
In case using “invisible” boxes , you mean not "mesh renderer"active? just the collider2D map that every “plane” or primitive 3D have? something like this ?
I’m suggesting to set up a 3D scene. With orthographic camera. Then put the map onto camera-aligned plane. Then setup invisible walls to prevent characters from walking where they aren’t suposed to walk.
Also look up “Saints of Range” on youtube and see how it is actually done:
Ok but, what’s about “z” axes? I mean I continue use x-y axes to move left and right direction and Up as Y for jump or moving Up-Down in my map? I prefer to work with entire 2D sprites, the configuration of my camera is 320x240 pixel in a 4:3 around aspect ration… the projection is orthografic … I need primitives shapes ,like boxe,planes ? to mimic and match the scenario ?
I got invisible plane for now , I can edit for suer for making more “long” … but probably I need to create another inviisible plane for the slope patch (pebble-stoned path in level).
if you want to make pixel - perfect stuff, never tried Unity’s 2D - stuff, however the Godot engine was built originally from a 2D - engine, and think has one of the best for pixel - perfect content . . .
in Unity, think overall people that make 2D, actually make what’s called 2.5D, or ’ faked-from-3D ’ effects, ie. what you’re working on now, however what you need to do is actually make the invisible walls, ie. a box - collider in the air, or go into Blender, make a custom shape, and think then make that a ’ mesh collider ', that way you could add weird slopes, or other stuff to the normally 100 % 2D - genre, it’s basically just the more normal 3D genre, where certain elements are forced into the 2D world // 2D controller, over a 3D - setting, or world . . .
good luck, think what you need is the ’ mesh collider ', also make some custom meshes in Blender, it’s also #technically easier to start with a 100 % 2D - system, etc etc . . . .
You can handle “Z axis” however you want. Your level could be flat. Or it could try to follow shape of every ramp and precisely position colliders to match it. Doesn’t matter. Both ways can be used.
Also instead of planes it is better to use boxes for collision.
good luck, think what you need is the ’ mesh collider ', also make some custom meshes in Blender, it’s also #technically easier to start with a 100 % 2D - system, etc etc . . . . [/QUOTE]
custom meshes in blender? you mean export the map as “!backcground” image? and start to modelling blocking such as this one? shapes? I can do it but it is so messy? I mean it can be different variation in z-blender axes, do not matter ,unity converts the z-axes to Y-axes automatically ,the problem is how to create that shapes? in the playable areas?and cover the limits as well…
You can handle “Z axis” however you want. Your level could be flat. Or it could try to follow shape of every ramp and precisely position colliders to match it. Doesn’t matter. Both ways can be used.
what do you think to use pro-builder? I know how to use blender ,but my auto-rig pro addon is not working property to export to unity very well… the default exporting feature from blender to unity is odd, messy and not work very often. So I discard blende for exporting the fbx… so I decided to create many blocks in pro-builder , the problem is I do not any idea how and where to put that blocks? I decided to put it similar like “saint range” game…
it’s also easier to make a simple 3D structure of the stuff, and then make the art fit later, try and make what the images look like, should be easier in the orthographic view, etc etc . . .
think to build the correct stuff in 3D first, in ’ theory ’ the 2D stuff should also fit, or it’s the ’ correct ’ places to walk, based on the images . . .
It doesn’t matter what you use. You could just place and scale a bunch of boxes in unity and it would work. You could use pro builder, you could use blender, you could use realtimeCsg/SabreCSG. Any method that works is usable.
Regarding autorig, you may need to redownload the addon from blender market. Those do not autoupdate and need to be udpated manually.
Also, autorig to unity export is definitely made for characters and not background geometry.
Here is the way… I ca not make it with blender cause the -90 axes rotation and I do not like the way blender handles the fbx … so I use builder pro here in unity , at the end of the day I need a mesh and colliders… but what’s about the “slope” I followed the shape similar that the other user suggest me.
This is pretty much how I would try to do it. Collision geometry is 3d, background is in the back, character sprite is aligned with camera. If there’s an element that should block the view, i’d extract it and place it in front of camera. Obviously the boxes would be invisible to the camera (it is possible to place them at invisible layer or disable renderers)
For example, you can place boxes into the scene. The boxes can be rotated and scaled. They have collision. You can assemble basic collision out of those boxes. without ever touching blender, probuilder or anything else.
Likewise, you do not have to precisely reconstruct geometry displayed on your background image. The floor can be flat, even if there are slopes painted in background.
