Introduction to Computer Graphics (Lecture 17): Rasterization

6.837: Introduction to Computer Graphics
Autumn 2020
Many slides courtesy past instructors of 6.837, notably Fredo Durand and Barbara Cutler.

[@ElSarcastro]

By random chance I stumbled upon this treasure trove of knowledge while pondering the difference between different rendering methods in Blener. I had several computer vision and computer graphics courses during my studies in 3 countries but never I was able to truly grasp the concepts. I wanted to express my gratitude for your hard work, effort and dedication in delivering this amazing series of lectures of superb quality

[@erich_l4644]

Thanks for the uploads! You're really good at speaking like an academic. Your explanations come out like finely worded prose.

[@bubu88134]

Thank you

[@GobalKrishnanV]

Need books related Rasterization. without using OpenGL or other Api's

[@stacksmasherninja7266]

It really must be tricky to implement portals ! I wonder how they did it so efficiently on some fairly old hardware

[@Yarogor]

I got difficulties with z-buffer understanding.
Am I understand it right, that z-buffer calculating constantly for every single pixel of EVERY segment in runtime?
Isn`t it mean that z-buffer contains depth values of All pixels of ALL segments(polygons) within view frustum simultaneously?
Or
Objects just drawing one by one, their pixels compare their Z value with current Z-buffer and just become visible or not?
If the second option is correct then I completely don`t understand how shadow mapping work :)

[@hjups]

The last option is correct, there was pseudo code at 23:00 which shows this. Only one depth value is stored for each pixel, and is updated sequentially for each intersecting triangle. However, each pixel can be updated in parallel.
For shadow mapping, effectively the pixel shader will do a 2-step ray trace for each pixel. Given the X,Y, and depth, you can calculate where in world space the rendered pixel is. Then you can find the distance from that pixel to the light source (with the shadow map). If the distance is greater than the depth value in the shadow map, then the pixel is in the shadow of that light source. Then you repeat for each light source (obviously that doesn't really scale well over many lights though).

[@Yarogor]

@@hjups THANK YOU! Now I got it )

[@grafix4955]

Nice 👍