This is the best explanation for window function I have ever come across.
I have one question: after changing the block size, the DFT doesn’t report the correct frequency content. In your example, there are 1, 3, and 4Hz signals, but after chopping the signal into 5 blocks, DFT reports additional signals at 0, 2, and 5 which windowing has minimum effect on them. Do we have ways to get around this problem? Thanks
Thanks for your kind words and question. I don’t know of a way of getting around the problem. Windowing reduces spectral leakage, but it doesn’t solve the problem completely. The problem is, by cutting the signal up, we are actually changing it. The best way of alleviating the problem is to cut a very long signal up into lots of very short sections. The number of sections needs to be very large and the length of each section very short when compared to the overall signal.
Would you like to learn about how the Fourier Series became the Fourier Transform, how the DFT and FFT work and what Windowing is? Please help me to devote more time to filming the rest of the video course by supporting me on Patreon.
Receive updates on the continuing production of the course, get notified when there is a new post or video and be the first to know when the full course goes live.
Unlock with Patreon
Great article.
Thank you.
Kudos, never seen such simplest and interactive explanation then this article on the topic. great going.
Thank you.
This is the best explanation for window function I have ever come across.
I have one question: after changing the block size, the DFT doesn’t report the correct frequency content. In your example, there are 1, 3, and 4Hz signals, but after chopping the signal into 5 blocks, DFT reports additional signals at 0, 2, and 5 which windowing has minimum effect on them. Do we have ways to get around this problem? Thanks
Thanks for your kind words and question. I don’t know of a way of getting around the problem. Windowing reduces spectral leakage, but it doesn’t solve the problem completely. The problem is, by cutting the signal up, we are actually changing it. The best way of alleviating the problem is to cut a very long signal up into lots of very short sections. The number of sections needs to be very large and the length of each section very short when compared to the overall signal.