Can an ideal box filter achieve perfect anti-aliasing? From the Nyquist Theorem we achieve no aliasing by sampling at twice the highest frequency. From this slide it seems that we can only limit the frequency range to the same as the sampling rate (1 pixel) through a box filter. So does that mean some aliasing still exists?
yinxudeng
@JefferyYC I think some aliasing will still exist because as we can see from the box filter in the frequency domain, it doesn't filter out higher frequencies completely.
yinxudeng
So far in this lecture, we have been talking about everything in terms of frequency. However, here we talk about box filter in terms of its width and attenuated frequencies in terms of its period. Is it possible to explain this filter and its attenuating effects in terms of frequency?
Can an ideal box filter achieve perfect anti-aliasing? From the Nyquist Theorem we achieve no aliasing by sampling at twice the highest frequency. From this slide it seems that we can only limit the frequency range to the same as the sampling rate (1 pixel) through a box filter. So does that mean some aliasing still exists?
@JefferyYC I think some aliasing will still exist because as we can see from the box filter in the frequency domain, it doesn't filter out higher frequencies completely.
So far in this lecture, we have been talking about everything in terms of frequency. However, here we talk about box filter in terms of its width and attenuated frequencies in terms of its period. Is it possible to explain this filter and its attenuating effects in terms of frequency?