The sampling errors make this plot look like some sort of trippy Gravitational Lensing effect f(z) = e^(i |z|⁵ )pic.twitter.com/FwI2TpFbf9
Show this thread
Visualizing computational limitations f(x,y) = sin(x²+y²) x,y∈(-200,200) The plot should be radially symmetric, though it only appears to be so around the originpic.twitter.com/axyIoO1mmu
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
you may want to try point sets with quasi Monte Carlo. For instance, a blue noise point set would drastically reduce aliasing artifacts. The aliasing you observe is mostly due to the Fourier spectrum of the sampling pattern used to integrate the function within each pixel.
-
i actually don't care about what the real plot looks like haha .. the real plot is actually rather boring .. thanks though
- 1 more reply
New conversation -
-
-
You can get a closed form expression for the average of f over each pixel, in terms of Fresnel integral functions. I’ve plotted that here for x,y∈(-20,20). As you’d expect, the integral of f over an axis-aligned square pixel isn’t a radially symmetric function.pic.twitter.com/VqXEVq9Y2U
-
Interesting : we see the (unwanted) circles way better on the Twitter previsualization than when we click on it.pic.twitter.com/38jja4UlXB
- 1 more reply
New conversation -
-
-
Nice. How come the aliasing results in a regular grid?pic.twitter.com/fvkkl1443m
-
Here's one with 4x subpixel resolution. There are still some aliasing effectspic.twitter.com/3YQHO3kOLQ
End of conversation
New conversation -
-
-
And here is another contour plot of the same functionpic.twitter.com/xWpSTLHfoD
- 1 more reply
New conversation -
-
-
Aliasing is art.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.