OK.

So welcome for computer graphics.

And as a reminder, or a new message for those who are not

here on Tuesday, so next Tuesday, December 3,

I think it should be, there will be no computer graphics

lecture, OK?

And we will compensate that by having two Thursday lectures,

which are 90 minutes instead of 45.

Yeah?

So next Tuesday, no lecture.

OK.

Good.

So on last Tuesday, we were speaking

about occlusion, visibility, and transparency.

And yeah, transparency is an interesting topic

because, as I told you or showed you last time,

it makes things more difficult. Transparency does not

work together with a set buffer.

So we explicitly have to take care of that.

And in standard renderings, like OpenGL or so, for instance,

we do not have a simple option that

says enable transparency.

And we can then render transparent objects.

But we really have to take care and look

at that when doing rendering.

And there are two reasons why transparency is difficult.

So the first one, that's just a repetition from last time.

The first thing is that alpha blending is not commutative.

Alpha blending means you combine,

when setting a pixel, you combine.

You do not just overwrite that pixel with a new color.

But you combine the old color with a new color.

And it's a linear combination.

And this linear combination value

is alpha, or alpha controls that linear combination.

And that can be interpreted as a transparency.

Or in fact, it's the other way around.

It's the opposite of transparency.

That's what we call opacity.

And by this, transparency can be achieved.

But transparency is not, or this alpha blending

is not commutative.

That means if I first render a red object

and then a transparent green one on top, it's different.

Then if I first render, or no, the other way around.

If I first render a green object and then

a red transparent on top, it's different from first rendering

a red one and then a green on top.

That's different.

It makes a difference whether I have a green glass in front