OK.

So hello and welcome for computer graphics.

We are right now speaking about advanced ray tracing.

And a method I presented last time

is called distribution ray tracing.

And the idea I explained for the example

that we want to render a material which

is not perfect specular.

And that means the reflection towards the viewer

does not only depend on one single reflection ray,

but in fact, it depends on a number

or on the entire region around the major reflection.

And now the problem with ray tracing

is that such a ray is always just a single sample.

It's just a single point.

Or it's not a point, but in a space, in a ray space,

it would be a single point.

So it only samples a certain direction.

But in fact, we would have to integrate

about this entire cone.

And the idea of distribution ray tracing now is very simple.

It just means if we have such a case,

do not cast a single ray, but instead cast several rays

and average their results.

And because that problem happens or appears quite often,

this glossy BRDF is just one instance

of the very same problem, there are also

other applications for this idea.

And the first one I spoke about is anti-aliasing.

And here the idea or the problem is

that the color of a single pixel does not only

depend on the color at certain points,

but it depends, it's really kind of the average color

over the entire pixel.

And now what we can do is that in order

to get nicer anti-aliased edges, for instance,

we can try to sample that range.

So we spoke about several methods.

The first one is a uniform grid over the pixel.

But we also saw other methods, for instance,

randomly sampling within the pixel.

And yeah, a few different methods

to how to generate these samples.

OK, now another point where this issue becomes interesting

is depth of field.

And now depth of field, the effect of depth of field

is known to each one of you.

So here we have a, I think that's also already

slightly blurred, it's just not very well visible.

But here we see a typical effect.