With a Python program I generated a sinusoid signal and I added to it Gaussian noise. Now I want to compute the optimal SNR by applying a matched filtering algorithm. Since the noise is white (at least I think it is, I did not make any assumption on its distribution in frequency), does it make sense to whiten the data?
$\begingroup$
$\endgroup$
6
-
$\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$– Community BotCommented Jan 14 at 17:55
-
$\begingroup$ You have analytical equations for the signal $S$ and for the noise $N$. The $S/N$ ratio has an analytical expression as well. What are you therefore expecting to vary to search for an "optimal" $S/N$? $\endgroup$– Jeffrey J WeimerCommented Jan 14 at 18:14
-
$\begingroup$ @JeffreyJWeimer Isn't in this way matched filtering work? I mean, the idea behind the algorithm is trying to maximize the SNR, at least this is what I understood $\endgroup$– AleNekro97Commented Jan 14 at 19:45
-
$\begingroup$ @Community I have done a program that generates a time series which is a sine, then I added a random Gaussian noise to it: When the noise is not white, in order to apply the matched filtering algorithm one has "whiten" the data. In this case noise is already white, so, does it make any senso to whiten the data? $\endgroup$– AleNekro97Commented Jan 14 at 19:48
-
1$\begingroup$ Are you sure this is physics-related question? I would ask that one on CrossValidated or Data Science Stack Exchange. $\endgroup$– Daigaku no BakuCommented Jan 14 at 19:57
|
Show 1 more comment