function [xsnap,Shat]=mksnaps(S,nsnaps)
%%% MKSNAPS
%%%   This function generates a set of snapshots given an ideal 
%%%   covariance matrix
%%%
%%%   Variables:       S = ideal covariance matrix (NxN)
%%%               nsnaps = # of desired snapshots
%%%                xsnap = matrix of snapshots
%%%                 Shat = estimated covariance matrix based on snapshots
%---------------------------------------------------------------------------
[nrcv,junk]=size(S);
[Sevec,Seval]=eig(S);

xsnap=Sevec*sqrt(abs(Seval))*(randn(nrcv,nsnaps)+j*randn(nrcv,nsnaps))/sqrt(2);
Shat=xsnap*xsnap'/nsnaps;
%---------------------------------------------------------------------------
return