how to calculate and move obj to top of Y on another object?

jbskaggs

This should be simple but its eluding me.

in mostly top down game I have a special effect explosions, splats, etc (animated3d sprites) that on my larger objects are too low on the y=axis and cant be seen.

What is the simplest way to calculate the height of the model (in current scale) and move the sfx object up the yaxis accordingly?

Or would it be easier to just unproject and use a 2d animations?

TwistedTwigleg

I haven't tried it myself, but if you can access the MeshInstance of the object, I think you can use the get_aabb function (documentation) to get the size of the mesh on the Y axis and then position it relative to that. Something like this:

var mesh_node = $MeshInstance
var effect_position = mesh_node.global_transform.origin
effect_position.y += mesh_node.get_aabb().size.y

Then I think it will be at the same height as the highest point as the mesh.

You could also potentially unproject and use 2D animations instead. What I think may be a little harder with that solution is if the camera moves, as then you would need to perform the unprojection code again so it's still in the right position on the screen. It should be doable though.

cybereality

You would ray cast from some plane directly above the world (say 10 units above the floor or whatever is larger than your largest object). Then shoot a ray straight down on the y axis. The first thing it hits is the y you need.

cybereality

https://docs.godotengine.org/en/stable/tutorials/physics/ray-casting.html

jbskaggs

I keep forgetting about raycasting.

xyz

It'd be simpler to draw these effects as 2d overlaid sprites. Because sooner or later you'll want to implement some camera movements and things may end up misaligned again.

Alternatively, you could place 3d sprites at exact 3d positions and just disable depth testing for them. They'll be drawn over everything.

jbskaggs

@xyz said: It'd be simpler to draw these effects as 2d overlaid sprites. Because sooner or later you'll want to implement some camera movements and things may end up misaligned again.

Alternatively, you could place 3d sprites at exact 3d positions and just disable depth testing for them. They'll be drawn over everything.

In this game the camera does not move, and I need simple as possible so I can try and finish it. :)

xyz

Both of the approaches I suggested are simpler than doing the raycast to determine 3d object boundary. They'll also look better because the sprite is always properly aligned, and there's practically no coding involved.

Positioning a sprite above the object in 3d will always look a bit off if view is not strictly top-down. The problem becomes more pronounced, larger the object because you need to push it more upwards. And you still may decide to lower the camera angle at some point for whatever reasons.

jbskaggs

What I got to work was create a 3d object for the FX, then add a 2d node under the Spatial node. then in script I added:

extends Spatial
    func _process(delta):
    	get_node('Node2D').global_position = get_viewport().get_camera().unproject_position(global_transform.origin)

This works but really I am not sure if I am implementing it properly.

cybereality

Yeah, I think the is the easiest way to do it.