Introduction to Computer Graphics (Lecture 19): Shadow maps, shadow volumes, deep shadow maps

[@nicholascunningham6936]

Love when a professor posts their lectures publically online because it shows that they actually have a desire to teach and spread their knowledge, not just get paid for their knowledge. Howdy from Texas A&M, really enjoyed the lecture!

[@HTMangaka]

The first time I started thinking about how this might work was when I first played Tom Clancy's Splinter Cell(The first one) back in 2003. I distinctly remember being like: "Huh. Most lights seem to cast a shadow. I notice that some are lower resolution, which implies that the shadow data is being stored in some kinda texture or matrice." I was 11 years old when I had these thoughts. 21 years later, I finally find out I was right. XD Stencil shadows explains Doom 3 SO MUCH. XD

[@AlphaGarg]

Fun fact about shadow volumes: they are still used in both the Enfusion (DayZ Standalone) and Real Virtuality (Arma 3) engines for opaque geometry like characters, vehicles, buildings, etc, while also using shadowmaps for non-opaque geometry like glass or foliage. Very nice results when it works!

[@Bilal-tu9rh]

16:00 The diagram and explanation are done well. Thank you 3000!

[@ITVMaudlin]

Great video! Explains something I've never bothered to figure out before quite clearly.

[@Yarogor]

The best lecture about shadow mapping! Thank you so much!

[@jonmichaelgalindo]

In VR, we can really get away without shadows! :-D We perceive depth and scene structure directly from stereo visual data. Still, they are awesome for realism / immersion, if your computation time can afford them. Thank you for the lecture!

[@blendernoob64]

I love the look of shadow volumes and would love to implement them into my unity game. Thanks for the lecture!

[@Iridium.]

Finnaly understood the depth comparison between the eye and depth texture . Thanks so much z

[@polpop4509]

this is a quite complete video about the topic, thank you very much for it!

[@firiasu]

Thank you so much! Very interesting and clear!

[@npip99]

Shadow volumes seem great with extremely low poly models. You can still have a 50k poly player, but have an "Ocarina of time"-poly player for shadow volumes. It seems like the pixel shader would be: void fragment() { stencil += sign(dot(normal, eye - position)); } And that'll populate the stencil with the total count. Very simple!

[@cahierdexercices9501]

Thank you so much for sharing your lectures

[@johnsports_iii]

It's kinda sad that shadow volumes aren't more efficient. Seems like a pretty neat algorithm.