I am a beginner of DSP. And these days I am confusing with the zero phase filtering in STFT as every segment after using the window will be zero padded and shifted to fulfill the zero phase. What's the advantage to realize the zero phase? As the purpose is to reconstruct the original signal using the inverse STFT, is the zero phase useful and key to get the original signal?Thank you!
Looking forward to your reply!
JOS wrote:
6/1/2010
If all you are doing is performing the STFT and reconstructing the original signal by an inverse STFT, then the phase (or even type) of the window doesn't matter. All that matters in that (degenerate) case is that the windows overlap and add to a constant at the chosen hop size R.
hmcherry wrote:
6/1/2010
Thank you for your answer.
I am still not totally understand what happens in the degenerate case if I don't add zeros and shift the window. Can you tell me the difference in detail?
If the original signal need be reconstructed, is the hop size R of some window is fixed?
JOS wrote:
6/4/2010
Well, I wrote down what I know in pretty good detail in this book! You might want to start with the previous book Mathematics of the DFT. You seem to be weak on basic Fourier theorems, so I would recommend getting to the point that you can prove all the main theorems for the DFT case.
Add a Comment
You need to login before you can post a comment (best way to prevent spam). ( Not a member? )