1
$\begingroup$

We have a signal $s(t)$.

  1. If we do an upsampling, does the signal duration increase?
  2. What is the point of upsampling if the signal time increases?
  3. Can we do an upsampling if we don't use a shaping filter?
$\endgroup$

1 Answer 1

1
$\begingroup$

No, if you upsample a signal, its duration (i.e. measured in real-world units like "seconds") does not change.

  • If your signal $s[n]$ is sampled at a rate $f_s$ Hz, and it is $N$ samples long, then the total duration of the signal is $\frac{N}{f_s}$ seconds.

  • Say you upsample the signal by a factor $M$, yielding a new signal $s_u[n]$. It will be $MN$ samples long, but it will be sampled at a new rate $Mf_s$. Its duration is therefore $\frac{MN}{Mf_s} = \frac{N}{f_s}$. You can see that the signal duration is the same as $s[n]$.

It's not clear what you mean by a "shaping filter" in your third question.

$\endgroup$
2
  • $\begingroup$ In the third question, 'shaping filter' is for example a root raised cosine filter: Can we do an upsampling if we don't use a shaping filter (for example a root raised cosine filter)? $\endgroup$
    – user39389
    Dec 8, 2018 at 19:25
  • $\begingroup$ upsampling has not necessarily something to do with pulse shaping, so yes. $\endgroup$ Dec 9, 2018 at 0:06

Your Answer

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

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