Can I think of the the reason why the edges are "signals with no band limit" as something like, since the image will be rendered by tons of pixels/squares, you can only get infinitely close to rendering a line with those squares but will never actually get a line?
sponge
It was breezed over really quickly in lecture so I didn't really understand why sinc would be associated with an n^2 bound. Could someone clarify?
tib
How would you apply the sinc filter to perfectly reconstruct the signal?
bpopeck
@UhrmasJHHue My interpretation was that edges can't be perfectly approximated by sine waves, no matter how high the frequency. Thinking about a horizontal edge ($y=0$), using a single sine wave to approximate it would require taking $lim_{t\to 0} \sin(t\cdot x)$, i.e. always increasing the frequency in order to make the wave more and more flat.
Can I think of the the reason why the edges are "signals with no band limit" as something like, since the image will be rendered by tons of pixels/squares, you can only get infinitely close to rendering a line with those squares but will never actually get a line?
It was breezed over really quickly in lecture so I didn't really understand why sinc would be associated with an n^2 bound. Could someone clarify?
How would you apply the sinc filter to perfectly reconstruct the signal?
@UhrmasJHHue My interpretation was that edges can't be perfectly approximated by sine waves, no matter how high the frequency. Thinking about a horizontal edge ($y=0$), using a single sine wave to approximate it would require taking $lim_{t\to 0} \sin(t\cdot x)$, i.e. always increasing the frequency in order to make the wave more and more flat.