A real band-pass signal $s(t)$ with center frequency $f_c$ can always be represented as

$$ s(t) = A(t) \cos(2\pi f_c t) $$

where $A(t)$ is real and can be viewed as amplitude.


$$ s(t) = A(t) \cos(2\pi f_ct) = \Re\left\{(A(t) e^{j2\pi f_ct})\right\}$$

Then we consider

$$s_l(t)= A(t) e^{j2\pi f_ct}$$

to be the corresponding complex band-pass signal of $s(t)$.

My question is how can I find the real low pass equivalent of $s(t)$ and the complex low pass equivalent of $s_l(t)$?

  • $\begingroup$ it's $A(t)$. for both (because $A(t)$ must be real because $s(t)$ is, then the Fourier transform of $A(t)$ (geez, what letter do i use if cap $A$ is the time-domain function?) is Hermitian Symmetric (i think that was the term). if you want to model $A(t)$ (and i would call it "$a(t)$" if i were doing this) as having an asymmetric spectrum, then i would make it complex for $s_l(t)$. $\endgroup$ – robert bristow-johnson Jul 25 '15 at 17:28
  • $\begingroup$ @robertbristow-johnson thanks, I would like to know one more thing though. Is there a process that I can follow to obtain the complex band pass signal from the real band pass signal? $\endgroup$ – Wanderer Jul 25 '15 at 17:39
  • $\begingroup$ yes. it's called the Analytic signal. you get it by adding your original real signal $s(t)$ with the Hilbert transform of it (times $j$ ). $\endgroup$ – robert bristow-johnson Jul 25 '15 at 17:45
  • $\begingroup$ @robertbristow-johnson Thanks again Robert. One last thing. Can any of the signals I mentioned above be characterized as low pass signals or as high pass signals? $\endgroup$ – Wanderer Jul 25 '15 at 18:09

It's not true that any real-valued band pass signal can be written as shown in your first equation:

$$s(t)=A(t)\cos(\omega_c t)\tag{1}$$

The most general representation of a real-valued band pass signal is given by

$$s(t)=\Re\{a(t)e^{j\omega_c t}\}\tag{2}$$

where $a(t)$ is a complex-valued function, called the complex envelop of $s(t)$. It is the complex baseband (low pass) representation of $s(t)$. With $a(t)=a_I(t)+ja_Q(t)$, (2) can be written as

$$s(t)=a_I(t)\cos(\omega_c t)-a_Q(t)\sin(\omega_c t)\tag{3}$$

where $a_I(t)$ and $a_Q(t)$ are called the in-phase and the quadrature components, respectively. Both $a_I(t)$ and $a_Q(t)$ are real-valued baseband (low pass) signals.

A third, equivalent, representation of $s(t)$ is obtained by writing the complex envelop as


where $b(t)$ is the (real-valued) envelop, and $\phi(t)$ is the phase. Using (4) and (2), the band pass signal $s(t)$ can be written as


Equations (2), (3) and (5) are three equivalent representations of a general real-valued band pass signal. If you compare representation (5) to Equation (1), you see that Equation (1) is a special case for $\phi(t)=2\pi k$, $k=0,1,\ldots$, i.e. for a constant phase which is zero (or a multiple of $2\pi$). From Equation (3), this is equivalent with the condition that the quadrature component $a_Q(t)$ is zero.

Note that the complex-valued band pass signal

$$\tilde{s}(t)=a(t)e^{j\omega_c t}\tag{6}$$

is the analytic signal mentioned in a comment. It has no negative frequency components and it can be obtained from $s(t)$ by computing


where $\mathcal{H}\{\cdot\}$ denotes the Hilbert transform.


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.