Blender 3D: Noob to Pro/The Empty Object

Every object in a scene has common settings in the Object Context. This includes the common transform settings shown at right: location in the scene, overall rotation and overall scaling. Most objects also have additional properties, such as the geometry of a mesh and the materials that govern how it appears in the render.

The Empty object has none of these additional properties. It has the overall transform settings, and not much else. And it has no appearance in the final render. So what is the point of having such a thing?

In fact, it has many uses, such as:

• In the Physics Context, it can be set as the source of a force field (e.g. wind).
• It can be used as the parent for multiple other objects. That way they can be moved together just by moving the Empty, rather than each child object individually.
• It can be used as a target object in an array modifier, where it adds its transformation to the copies of the object being modified by the array.
• It is a handy marker and place holder if you are moving things around. You can mark an object's original position, and then put it back easily where the empty object is.

In these situations, the invisibility of the Empty in the render is an advantage, because it can be placed wherever necessary, without introducing unwanted clutter into the final image.

Using an Empty With the Array Modifier

The types of Empties

Open a new default Blender document, complete with default cube. Add a new Empty object to the scene. It doesn’t matter what type of Empty you choose; the options just control its appearance in the 3D view, not the actual behaviour of the Empty. Besides, you can change your mind about this later anyway.

Now select the default cube, go to the Modifiers Context and add an Array modifier. In the properties header, there are three different ways the modifier can produce its copies of the original object: via a Constant Offset, Relative Offset, or Object Offset. The Constant and Relative offsets only allow a simple displacement for the copies, but the Object Offset looks at the full transformation of the specified object, which can include rotation and scaling.

Here I have unchecked the default Relative Offset, checked Object Offset, and from the popup menu for selecting the target object, I have selected my Empty (which by default would have been called “Empty”).

But if you look at your 3D view, nothing seems to have happened: the cube is still sitting, unmoved, in its original position. In fact, it might be hard to see the Empty, because it would have been inserted in the same position as the cube, so it ends up within it. If you switch to Wireframe  Z  view, you should be able to see the Empty sitting within the cube.

Before doing anything further, increase the Count field in the cube’s array modifier to 3. This will give you a clearer idea of the effect of subsequent manipulations. Doing so will not have any noticeable effect to start with, because all the copies will be lying right on top of each other.

Now select the Empty, and move it to one side: you will immediately see a second copy of the cube follow the empty, and the third copy move by the same amount in the same direction, so it ends up beyond that.

Return the Empty to its original location, and this time try moving the cube. As the original cube moves, you will see the second copy stay with the Empty at the centre of the scene, while the third copy moves beyond that, by the same distance and in the same direction as the second copy is removed from the original cube.

Now if you return the cube to the same location as the Empty, select both and try moving them together, you will see no additional copies of the cube appear, because they remain exactly on top of the original cube.

The rule for using an Object Offset with the Array modifier is this:

The transformation applied to copies of the modified object is the difference between the transformations of the target object and the modified object.

In other words:

• the translation applied is the difference between their origins.
• the rotation applied is the difference between their object rotations.
• the scaling applied is the ratio between their object scaling factors.

Try moving the Empty to one side, as before; this time leave it there, so the three copies of the cube are nicely spread out. Now try applying a rotation to the Empty; see how this applies the corresponding rotation to the second copy (the one located where the Empty is), while the third copy gets twice that rotation.

Return the Empty to its unrotated state, and this time try rotating the cube: see how the second copy (where the Empty is) stays unmoved, while the third copy gets the opposite rotation.

Undo the rotations, and now try scaling: shrinking/enlarging the Empty correspondingly shrinks/enlarges the second copy, while the third copy gets transformed by the square of the shrink/enlarge ratio (the ratio multiplied by itself, or alternatively the ratio raised to the second power). Or shrink/enlarge the original cube, and see the second copy stay unchanged (as before), while the third copy enlarges/shrinks by the inverse ratio.

When you change the transformation of the original cube, the second copy always stays unchanged because the Object Offset transformation is the inverse of the transformation you are applying to the original; hence it always cancels out for the second copy, but you can see corresponding increasing powers of it applied to subsequent copies.

Editing the Modified Object

Make sure the Empty is positioned so the array instances are well separated. Select the cube,  TAB  into Edit mode, select all the vertices, and try transforming them: moving, rotating or scaling.

Note the difference in behaviour from doing this to the cube in Object mode: this time all the copies of the cube transform along in unison. This is because we are not affecting the object transformation, which is what controls the behaviour of the array modifier.

Arranging Object Copies in a Circle

Which brings us to a convenient trick for a common need: arranging copies of an object in a neat circle.

The easiest way to do this is start with the cube and the Empty in their original central positions. Suppose you want 8 copies of the cube in your circle; the angle between them then needs to be 360° ÷ 8 = 45°. So rotate the Empty by 45°. Now select the cube. Set the Fixed Count field of its Array modifier to 8.  TAB  into Edit mode, and move all the vertices to one side. You should see the other 7 copies of the cube correspondingly move away from the centre in different directions, maintaining a nice neat circle arrangement. Your cube mesh is simply being rotated about its origin, which is not where the vertices are, but where the Empty is!