0
$\begingroup$

Let assume we have a signal as:
$\color{blue} {x(t)=M e^{(-\beta.t)} e^{j(\sigma .t+\delta )}}$
By applying phase angle operator we have:
$\color{blue} {Phase[x(t)]=Phase[M e^{-\beta.t} e^{j(\sigma .t+\delta )}]=\sigma.t+\delta}$
I have two questions:
1. How can I implement the Phase angle operator, i.e. Phase, in practice ,like a chip.
2. Does phase angle operator increase the effect of noise in a noisy signal?(for example some operators like derivative operator increases the noise effect in a noisy signal)
Kind Regards

$\endgroup$
  • 1
    $\begingroup$ I guess $\tau$ on the right-hand side should be $t$. If you already have a complex-valued signal (i.e. you don't need to generate an analytic signal from a real-valued signal), then the phase is simply the phase angle of the complex number represented by the signal at a given time $t$. So you basically need to implement atan2. $\endgroup$ – Matt L. Sep 22 '14 at 13:27
  • $\begingroup$ Yes, I meant to present the time variable by writing $\tau$ However the signal is modified and 't' is substituted now. Also, Thanks for giving the hint of using atan2. Now I have two questions. First, How can I apply it to my signal? is this function within every microcomputer programming tool? Second, Does noise effect increase the error as derivative operator does? thanks @MattL. $\endgroup$ – Electricman Sep 22 '14 at 13:49
1
$\begingroup$

As mentioned in my comment, you need the atan2 function. If it is not implemented in the library you're using, you can implement it yourself quite easily according to this definition. If your signal is $x(t)=x_r(t)+jx_i(t)$, where $x_r(t)$ and $x_i(t)$ are the real and imaginary parts, respectively, then for a given time $t$, the phase is computed as $\phi(t)=\text{atan2}(x_i(t),x_r(t))$.

The computation of the phase of a signal is indeed sensitive to noise. Note that - no matter how you implement it - there is always some direct or approximate computation of the ratio of the imaginary and the real part of the signal, and since both are corrupted by noise, the ratio can be very noisy. The computation of the ratio can also become numerically problematic if the denominator (i.e. the real part) becomes very small.

$\endgroup$

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.