OK.

So here we are finally after my computer booted twice

and installed updates in between.

Sorry for that.

I finally made it.

So what we will speak about today is modeling.

And I already gave a very, very brief introduction

on how to represent triangle meshes, for instance,

in the introduction of that lecture.

Because obviously you need some models in order

to be able to render them.

And today we want to look into that further.

We will look at different representation

of triangle meshes and stuff like that.

So yeah, what's the goal?

Most of you did programming exercises,

so you know the problem well.

Before you can render a scene, you

need some description of the objects.

And the typical way we do that is using triangle meshes.

There are other ways.

So for instance, in the upper left, what you can see

is a volumetric data set.

This does not have surfaces.

Instead, it's a kind of three-dimensional texture

that stores transparency values, as they typically

come out of a computed tomography machine

or things like that.

But the typical way, or a very often used way,

to render such volumetric objects

is, again, to transfer them or to transform them

to some geometry, to some surface that you can render.

And now these objects can be quite simple

as those on the lower left.

Or they can be much more complicated,

like the ones on the right.

They can be really modeled, or they

can be generated by a scanner or whatsoever.

So in the very first computer graphics demos,

people typically used very simple objects,

like a cube, a cylinder, or I think on the right,

that's an icosahedron.

So these are very simple objects in the sense

that they can be described using a few parameters.

A cube, for instance, can be described only

by its size.

So this is a very simple example.

A cube, for instance, can be described only by its side

length or a cylinder, by the radius and its height,

and the icosahedron just by the radius of the bounding sphere,