After a long first experiment, this one was shorter. In this we had two stages. First we converted x(n) to X(K) and plotted the magnitude spectrum. In the second stage, we zero padded the input with 4 zeros and saw the change in the magnitude spectrum.
We also did IDFT of 4 point signal to verify to get back the original input.
As the length of the N increased due to zero padding, frequency spacing of magnitude spectrum reduced and approximation error also decreased. With zero padding, the length of signal increases. So the resolution of the spectrum increases and hence the the approximation error reduces.
Although it was shorter than the previous one, it wasnt very easy to implement.
Codes can found at:-https://drive.google.com/open?id=0Bzfvoo_rjoa8S19TN2V1SE9kckk
As N increases, the number of points per unit of spectrum increases.
ReplyDeleteyes,appending artificial zeros to the signal, we obtain a denser frequency grid when applying the DFT.
Deleteas we increases samples, quality of of spectrum increases.
ReplyDelete