2
$\begingroup$

I'm stuck at exercise 1.18 in Understanding Digital Signal Processing by Richard G. Lyons:

There is an often-used process in DSP called decimation, and in that process we retain some samples of an $x(n)$ input sequence and discard other $x(n)$ samples. Decimation by a factor of two can be described algebraically by $$y(m) = x(2n) \tag{P1-3}$$ where index $m = 0,1,2,3,. . .$ The decimation defined by Eq. (P1-3) means that $y(m)$ is equal to alternate samples (every other sample) of $x(n)$. For example: $y(0) = x(0)$, $y(1) = x(2)$, $y(2) = x(4)$, $y(3) = x(6)$, $. . .$ and so on. Here is the question: Is that decimation process time invariant?

Illustrate your answer by decimating a simple sinusoidal $x(n)$ time-domain sequence by a factor of two to obtain $y(m)$. Next,create a shifted-by-one-sample version of $x(n)$ and call it $x_{shift}(n)$. That new sequence is defined by (P1-4):

$$x_{shift}(n) = x(n+1) \tag{P1-4}$$ Finally, decimate $x_{shift}(n)$ according to Eq. (P1-3) to obtain $y_{shift}(m)$. The decimation process is time invariant if $y_{shift}(m)$ is equal to a time-shifted version of $y(m)$. That is, decimation is time invariant if $$y_{shift}(m) = y(m+1)$$

I'm somewhat confused by the notation $y(m) = x(2n)$--according to the definition of the decimation, I would hve written $y(n) = x(2n)$. Anyway, if I try to resolve the problem as it is stated:

  1. "Illustrate your answer by decimating a simple sinusoidal 𝑥(𝑛) time-domain sequence by a factor of two to obtain 𝑦(𝑚)"

    Let say $x(n)=sin(\omega n) \text{ with } \omega=2\pi f_0 t_s$ $\implies y(m) = x(2n)=sin(2\omega n)$

  2. "create a shifted-by-one-sample version of $x(n)$ and call it $x_{shift}(n)$"

    $x_{shift}(n) = x(n+1) = sin(\omega n+\omega)$

  3. "decimate $x_{shift}(n)$ according to Eq. (P1-3) to obtain $y_{shift}(m)$"

    $y_{shift}(m) = x_{shift}(2n) = x(2n+1) = sin(2\omega n +\omega)$

  4. "That is, decimation is time invariant if $y_{shift}(m) = y(m+1)$"

    $y(m+1) = x(\text{???})$

Eq (P1-3) doesn't really give a relation between $m$ and $n$. So what is $m+1$? According to the decimation definition given in the exercice, I would be tempted to say that $m+1=2n+2$. But that leads to $y(m+1)=sin(2\omega n + 2\omega) \neq y_{shift}(m)$

Could you pinpoint my mistake or clarify my misunderstanding in this exercise?

$\endgroup$
4
  • $\begingroup$ Simple answer: when you shift $x[n]$ by $M$, $y[n]$ = shifts by $M/2$. That's not the same shift, so is violates the invariance condition $\endgroup$
    – Hilmar
    Dec 7, 2019 at 15:10
  • $\begingroup$ Thank you @Hilmar. As you see, I do my homework ;) Joke aside, that was mostly what I was writing in a comment below: $y_{shift}(m) \neq y(m+1)$ is the right conclusion. I don't know why I was persuaded we had to prove it was otherwise. $\endgroup$ Dec 7, 2019 at 15:15
  • 1
    $\begingroup$ You should indeed be confused by the notation, because the notation is wrong. Decimation by a factor $M$ is described by $y[n]=x[Mn]$. Please browse this site for book recommendations. In my opinion there are much better (and free) books available, such as the one by Orfanidis. $\endgroup$
    – Matt L.
    Dec 7, 2019 at 15:15
  • $\begingroup$ Thanks, @Matt. I will do that. That being said, the Lyons book is very accessible at my level. It refreshes concepts I studied a long time ago and introduce new ones progressively. I saw the post you mentioned and already downloaded a couple of free books from the list. I will double-check to see if the one by Orfanidis was part of them. $\endgroup$ Dec 7, 2019 at 15:20

1 Answer 1

3
$\begingroup$

Sample rate conversion systems (expansion or decimation) are time-varying operators.

For your example system of decimation by $M$ :

$$ y[n] = T\{x[n]\} = x[Mn] $$

you can easily see that results of shift in the input and output are not the same; i.e.,

$$ y_1[n] = T\{ x[n-d] \} = x[Mn - d] $$

and

$$ y_d[n] = y[n-d] = x[M(n - d)] \neq y_1[n]$$

$\endgroup$
2
  • $\begingroup$ The books I'm reading right now aren't using (yet?) the $T\{...\}$ notation> But, as a conclusion, and to use that particular exercise wordings, I was right to find that $y_{shift}(m) \neq y(m+1)$--so I don't need to spend more time trying to prove decimation is time-invariant, since this is the exact opposite... $\endgroup$ Dec 7, 2019 at 15:09
  • 1
    $\begingroup$ The system notation $y[n] = T\{x[n]\}$ is just a depiction that an input $x[n]$ is mapped (or transformed) into an output $y[n]$, nothing special ;-). Also your conflict was correct as the notation $y[m] = x[Mn]$ is misleading (or may be just a typo)... Time invariance is proved if you can show that the shifted output $y[n-d]$ equals the response for the equally shifted input $T\{x[n-d]\}$. $\endgroup$
    – Fat32
    Dec 7, 2019 at 19:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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