You are viewing the course site for a past offering of this course. The current offering may be found here.
Lecture 3: Sampling and Aliasing (83)
keatonfs

What is considered aliasing or not? To me, the 5th image still looks "jaggy". Is this just because of the dimensions of the photo?

neekrom

@keatonfs I think the reason you see the 5th image as jaggy is because of the resolution. In this new image, we took samples for a pixel every 16 pixels in the original image, so the resolution is lower. But it should still be clear that as you blur more you get less of that aliasing staircase effect on the edges of objects, like on the border between the girl's dress and the background.

eric-li-2

If we anti-alias an already anti-aliased image at a different cutoff frequency what happens? I mean a photograph is already anti-aliased by the hardware of the camera since it returns a grid of pixels. The camera essentially performed a one pixel convolution. Is there some way to use this fact to improve our anti-aliasing? Currently it feels like we are translating languages and using the translation of a translation. Any thoughts on this?

I feel like this is composing kernels, but I fear the resulting kernel may not have great properties.

You must be enrolled in the course to comment