Camera controls: Zoom vs. Dolly

The difference between zooming and dollying the camera is somewhat important to understand when working in 3D applications.

Scrolling the mouse wheel in Unity's sceneview applies the "Dolly" effect to the camera, not Zoom.

Dolly, basically means that the camera moves (translates) through the 3d environment and therefore, at some point, it will be positioned beyond the grid which is then invisible.

Zoom narrows the view-cone of the camera, which makes objects seem closer to the camera.

Dolly and Zoom is commonly mistaken for one another. The easiest way to visually spot these from one another is to look at the perspective. Narrowing the view-cone will decrease the perception of perspective, where dolly keeps it intact.

Why are these terms used interchangeable when working with 3D graphics?

I think the reason they're used interchangeably is largely down to lack of knowledge of the term 'dolly' among game developers - wheras it's a term much more commonly known in the film industry. As a result, the term 'zoom' is often mistakenly applied to both types of effect in games, and from a non-technical point of view, the word "zoom" sounds like it is the correct term.

I think it's important that you raise this point as I quite often find that people who are asking "How can I make my camera zoom in and out?" are actually after a dolly effect, rather than just a narrowing of the field of view.

I often describe it like this:

"Zooming-in" is effectively the same as simply magnifying a portion of the view as if it were a flat image. As a result, the more you zoom in, the flatter the perspective seems within the area of view that remains - think 'sniper scope view' and you have the right idea.

"Dollying-in" (which is often what game developers want when they say 'zoom') actually moves the camera towards the object. On its way, it may actually pass nearby objects. This is much more like actually flying towards your target to get a close-up view. This kind of camera motion is usually used when switching camera views in a racing game, for example where you might change between a further-away tail view and a closer-up tail view.

I think people use zoom because zoom has become commonplace in 2D application interfaces in which being 2D and orthographic the difference is moot. For a neat effect zoom in and dolly out at the same time. Very Spielbergian.