kirupa.com - 3D Movement

Question of the Week

What non-geek hobbies do you have?

	Flash
	Basics / Animation
	ActionScript
	Special Effects
	HTML5
	CSS / HTML Basics
	Animation
	JavaScript
	Common Tasks
	Windows Phone
	Get Help

3D Movement
by senocular

Exciting eh? Well there's really not much going on here other than adding a new span of heights to the existing movement in an isometric scene. The hard part has already been done - that's the movement within the skewed isometric surface. This new 3D angle is straight up and not skewed at all. Its the 'normal' one in the bunch. This makes our job a lot easier. To use this new dimension all you need to do shoot up your object at a real distance of whatever you want. Calculations not required.

Lets take a look back at the iso cube:

Confusion Alert!

Before, in moving in 2D space, we used x and y to represent positioning. This seemed completely logical since we are used to dealing with that in all other terms. Here, however, when adding in this 3rd dimension in the isometric world, we're adding a dimension which corresponds directly with _y.

The thing is, our concept of isometric y is already taken up in one of the original 2 dimensions. This third dimension, aka 'z', is not your normal z of back in depth, but your normal y instead. Because of that, confusion can occur. The best thing to do is think of the isometric space as looking straight down on it. Then this z axis is actually going in with depth and not up as y should.

With that being said. What does it take to get up there in that 3rd dimension? Well, only a change in a movie clip's _y. Want the clip to move vertically in space 10 places, add 10 to its _y and you're good! As easy as that seems, sadly, it doesn't stop there. What you need to do aside from that is also keep track of your z movement. Otherwise, what's to prevent you from thinking your object just moved diagonally in your original isometric plane?

You can see here that these two balls are physically at the same height and it could be assumed the blue ball lies on the square directly left of the red ball though the dotted lines indicate that blue ball is in fact closer to you and elevated into 3D space. So to keep track of this, we need to include in an isometric objects properties a z property to distinguish its position in the z or vertical space. Positioning the object is just a matter of positioning normally with no z positioning and just uping the _y of the object by z afterwards to give it the height.

In terms of code, not much is added or changed at this point aside from adding the z property in to objects in the isometric space. This will also be added to the point class so a point will represent not only x and y dimensions but also z. Other operations, however, don't have to really be altered so much (aside from just throwing in a z property) since the nature of this z dimension is just that which was described above; only a matter of adjusting _y post-normal movement.

So here's an example in iso-3D:

[ click play to see more ]

Along with what was described above, added to this example is some simple collision detection in the 3D space too. All that entails is just keeping references of those cube's locations in a 3D array (array for each dimension) and just seeing which isometric grid spot the ball is in and checking to see if the ball is in the same location as a cube. If so, the cube clip is played displaying a shaking animation.

Depth swapping here, btw, is based on the previous explanation in 2D. This is because no matter what the height of an object in this 3D space, its closeness to you will never change so its arrangement in the space doesn't really change. The problem, however, as previously mentioned, is that once you have a lot of clips on the screen, they'll start trying to share the same depth swapping each other out of their calculated depths and into their last.

This is because the same depth is obtained for a whole plane of movement along left and right and up and down. Methods of handling that can be discussed later. I think the next installment will be about actually putting all this crap to work in a fully functional isometric example. Theory and mumbo jumbo is one thing, but making it work is another.