I am solving a question that says if we have sequence $x(n)$ of a signal missing one term then we have to find autocorrelation matrix $R_x$ as follows:

$$R_x = E\{\mathbf {xx^H}\}$$ Now if I take $x(n) = [x(0) , x(2),x(3)]$ that is missing term $x(1)$ I found following value of $R_x$

$R_x = $$\left( \begin{matrix} r_x(0) & r_x^*(2) & r_x^*(3) \\ r_x(2) & r_x(0) & r_x^*(1) \\ r_x(3) & r_x(1) & r_x(0) \\ \end{matrix} \right) $

Now I check that this matrix is Toeplitz or not, by the definition of Toeplitz (constant diagonal term) $R_x$ is Toeplitz. But when I checked the solution it says the above matrix is not Toeplitz and there is no explanation given. I wonder what I am missing in definition of a Toeplitz matrix that does not fit in the above matrix. Can anyone explain this?


For $\boldsymbol{R}_x$ to be toeplitz it should be:

$$\boldsymbol{R}_x = \begin{bmatrix}r_{x}(0) & r_{x}(2) & r_{x}(3) \\ r_{x}(2) & r_{x}(0) & r_{x}(2) \\ r_{x}(3) & r_{x}(2) & r_{x}(0) \end{bmatrix} $$

Since it is not, it is not a toeplitz matrix.

  • $\begingroup$ you mean when they say diagonal should be constant means all diagonals should be constants? $\endgroup$ – Tab May 7 '16 at 3:12
  • $\begingroup$ Basically all the diagonal elements should be the same. If this is what you mean by constant, then yes. $\endgroup$ – Amal May 7 '16 at 3:16
  • $\begingroup$ oops yes I checked definition again it was just my lack of attention while reading it. $\endgroup$ – Tab May 7 '16 at 3:16
  • $\begingroup$ I understand. Been there multiple times myself. $\endgroup$ – Amal May 7 '16 at 3:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.