Rendering Lecture 5 - Monte Carlo Integration III

This lecture belongs to the computer graphics rendering course at TU Wien. We explain how to use Multiple Importance Sampling (MIS) to reduce the variance of a Monte Carlo estimator.

[@karpikvisuals]

Anyway, thanks so much for make this materials public. Its a huge help for CG community!

[@HardCoreCodin]

The content is very well done, but honestly, I really don't understand why you chose to teach this in that order - first you explain HOW to do something complicated without even covering what you intent to use it for and why - and only THEN you connect it back to the reason people are watching this in the first place? What's the logic there?

[@cgtuwien]

Hi, thanks for the critique :)
We will restructure some parts of the lecture for the second iteration. Obviously we want to make it more logical.
Could you elaborate on what you mean? Are you referring to this video in particular or the lecture in general? What exactly do you mean by 'something'?
Cheers, Adam

[@HardCoreCodin]

@@cgtuwien I mean that the next video after this is the rendering equation - the impossible integral that can not be integrated - which is the very reason for all these numerical approximation schemes. Before students know what it is, and why it can't be integrated analytically - why should they even care about any of these approximation/simulation business and all their mechanical details? Why should any of that even appear as relevant? It's walking through the solution, before even covering the problem. It's delving way into the deep ends of the answer, before even presenting the question. Hope that clarifies what I meant by that.

[@cgtuwien]

​@hardcore codin thanks. yes it does.
we do cover integrals over the hemisphere (direct lighting and ambient occlusion) already before. other reasons for this order are the practical coding exercises that are going on in parallel and that we wanted to alternate topics between more practical and more theoretical.
that being said, we'll try to include your feedback in the next iteration. thanks again and cheers from Vienna
Adam