I specifically have a problem with certain notation used to describe systems of signals (context: Signals and Systems by Oppenheim, et. al.); from the Center for Computer Research in Music and Acoustics (Stanford University):

enter image description here


The current material I am learning with likes to use something similar along these lines. Particularly, in proofs, I have a hard time differentiating between the system proper and the inputs to the system, as the system is not allocated a specific variable to represent it before evaluation/substitution/abstraction or "binding."

Can I represent a system with a more function-like notation: e.g.:

$$H[f(x)] = f(2x)$$

To me, this seems a lot clearer, but it seems to break down in the context of ordinary differential equations:

$$\frac{dy(t)}{dt} + (sin\ t)(y(t)) = 3x(t)$$

Would the following be a good notation for the above?

$$H[y(t), x(t)] = \{ \frac{dy(t)}{dt} + (sin\ t)(y(t)) = 3x(t) \}$$

Or, am I misunderstanding something that is causing me to "not get" the notation?

I am looking for a clear notation that expresses the signals and their dependencies clearly.


I am not sure what you exactly mean.

In general, $x[n]$ represents the input of the system, and y[n] the output. (Except for the specific illustration)

We always use $H(\cdot)$ to represent a filter. ( But you can use it to represent a general transform , just the notation)

If you define the function $H(\cdot)$ with the first argument $x[n]$ and the second one $y[n]$, then it seems no problem.

An advise: if you want to ask the questions about Signal Processing, please post question on this page :)

| cite | improve this answer | |
  • $\begingroup$ Is there a difference between a filter/transform and a system? I once asked whether the Fourier transform is a system, but I couldn't understand the answer... $\endgroup$ – CinchBlue Sep 23 '16 at 9:05
  • $\begingroup$ @VermillionAzure The difference bewteen the output and input is the transform( system transfers $x[n]$ to $y[n]$) . Filter is one of the transforms $\endgroup$ – stander Qiu Sep 23 '16 at 9:08
  • $\begingroup$ Can you restate that? It's not clear to me what you're trying to say. $\endgroup$ – CinchBlue Sep 23 '16 at 9:11
  • $\begingroup$ Fourier Transform is not a system. It does not transfers the $x[n]$ to other things. It is a mathematical tool to calculate the spectrum, which is another angle to look at the signal $\endgroup$ – stander Qiu Sep 23 '16 at 9:12
  • $\begingroup$ @VermillionAzure I am very sorry that I misunderstand the 'transfer'.( not native speaker) . Transform means changes. $\endgroup$ – stander Qiu Sep 23 '16 at 9:15

The problem is that in practice a linear systems is not "naturally" given by an explicit function $H(x) = y$ that maps an input signal $x$ to an output signal $y$, but rather an implicit relation like your example: $$ y'(t) + \sin t \, y(t) = 3 x(t), \quad y(0) = y_0 \quad (1) $$ Provided this equation has a unique solution, this equation represents a linear system $H$ that maps an input signal $x$ to an output $y$, but we have no explicit formula for $H$ in this case.

To answer your question, I think the best notation is to stick with (1), and simply state which signal is the input and which signal is the output if you need to be extra clear.

| cite | improve this answer | |
  • $\begingroup$ One problem I have with the notation is that having to specify what the input/output signal per "instantiation" or specialization of the general equation to specific signals (e.g. x1 into x and y1 into y) is that we lose the information that specialization of the general equation is derived from that same system. Also, isn't H in this case simply the entirely of (1) minus the initial condition definition? Or does this cause a problem with substitution? $\endgroup$ – CinchBlue Sep 23 '16 at 9:22

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.