Depth-buffer visibility is used for properly rendering a scene so that only pixels which are in front of the existing pixel are drawn. Any pixel which lies behind the existing pixel is not shown on screen. This checking is performed using a depth buffer which is a 2-dimensional array of depth or z values which map 1 to 1 to the frame-buffer. Normally modern games use (24-bit Z/8-bit stencil) or (32bit) z buffers. So you have 1 depth value for every pixel in the frame buffer.

Back-face removal is a performance enhancing optimization in which the faces whose normals are opposite to the direction of the camera are not drawn i.e. culled. This way they don’t need to be textured/lit/rendered.

You will have to sort the polygons as per their depth values if you want to know their visibility prior to sending them to the GPU for rendering.

Between the two techniques, can you compare their efficiencies or when one might want to use one or rather than the other?

This seems like a homework question, we are not here to do your work for you, though we are willing to help out. Why not take a shot at giving us an answer and we can guide you through more research or correct mistakes?

You don't have to answer if you don't know the answer. The above description is my answer to an old question. It is big enough for anyone to answer the below question.

While I can see where your impression can be taken, I'd say I have to disagree.

Generally speaking, the question is a bit over-researched to be a homework assignment, the work is practically done.

Note to previous answerers: If you don't know what you are talking about, please do not answer. This question makes perfect sense and is perfectly reasonable if you know anything about Computer Graphics.

I wasn't aware I had even attempted to answer a question, but thanks none the less for pointing out my wrong doing.

Ok, you DID realize that WW is a moderator? The site has a policy not to do peoples homework for them. You should take that to heart else you will not find this a very friendly forum.

Now on to your question:

Well I was a big fan of z-buffering -- it is a fairly strait forward way of dealing with the problem of rendering a scene. It DOES tend to lead to poor edge definition (z-fighting). You have to remember that it is actually a false 3D... it is a set of planes really.

Back Face removal on the other hand is more 3D friendly. The biggest problem is that it only works with convex polygons (well it can do some concave polygons just not all). For this reason you need to convert all of your shapes to collections of convex polygons.

You can read up on these techniques on sigraph.org (great place for info on computer graphics)

As for why you choose one over the other depends on what you are doing.