I'm sure many people have seen this type of artifact when playing video games at low resolution. Video games typically use rasterization because it's very cheap!
TonyLianLong
Jaggies are artifacts from rasterization, often from aliasing effects that causes high frequency components to appear (abrupt changes in brightness values). More about jaggies could be found on Wikipedia: https://en.wikipedia.org/wiki/Jaggies
austinapatel
From EE120 we learned about the Nyquist sampling rate, which is the lowest rate you have to sample a signal to get all the information from the signal. I wonder if Fourier transforms play a role in removing high frequency elements before rasterization to make jaggies less of an issue.
Staffmcallisterdavid
@austinapatel Wait for next lecture, but great intuitions!
As a follow-up, what would would you do with the Fourier transform of an image to remove high frequencies?
Sicheng-Pan
It seems that the jaggies can be noticed even if individual pixels cannot be distinguished when we use simple point sampling. So simply increasing the resolution may not be a good solution to the problem.
ShaamerKumar
@david cool question! please let me know if im incorrect, but based on the next lectures slides i think we first apply the fourier transform, to convert it to frequency domain. Then, we find the locations of high frequency and use a low-pass filter to remove them. We then use inverse of the fourier transform to convert it back and were done!
I'm sure many people have seen this type of artifact when playing video games at low resolution. Video games typically use rasterization because it's very cheap!
Jaggies are artifacts from rasterization, often from aliasing effects that causes high frequency components to appear (abrupt changes in brightness values). More about jaggies could be found on Wikipedia: https://en.wikipedia.org/wiki/Jaggies
From EE120 we learned about the Nyquist sampling rate, which is the lowest rate you have to sample a signal to get all the information from the signal. I wonder if Fourier transforms play a role in removing high frequency elements before rasterization to make jaggies less of an issue.
@austinapatel Wait for next lecture, but great intuitions!
As a follow-up, what would would you do with the Fourier transform of an image to remove high frequencies?
It seems that the jaggies can be noticed even if individual pixels cannot be distinguished when we use simple point sampling. So simply increasing the resolution may not be a good solution to the problem.
@david cool question! please let me know if im incorrect, but based on the next lectures slides i think we first apply the fourier transform, to convert it to frequency domain. Then, we find the locations of high frequency and use a low-pass filter to remove them. We then use inverse of the fourier transform to convert it back and were done!