Matlab for the DPSS Window

The function dpss in the Matlab Signal Processing Tool Box^4.10 can be used to compute them as follows:

  w = dpss(M,alpha,1); % discrete prolate spheroidal sequence

where $M$ is the desired window length, and $\alpha\in[0,M/2)$ can be interpreted as half of the window's time-bandwidth product $\Delta t\cdot \Delta f$ in ``cycles''. Alternatively, $\alpha$ can be interpreted as the highest bin number $k$ inside the main lobe of the window transform $W(\omega_k)=\hbox{\sc DFT}_M(w)$ , when the DFT length is equal to the window length $M$ . (See the next section on the Kaiser window for more on this point.)

Some examples of DPSS windows and window transforms are given in Fig.3.29.