0
$\begingroup$

Some methods

#1

YData = YData / max(abs(YData));

#2

maxYdataAbs = abs(max(YData));
minYdataAbs = abs(min(YData));
YData = YData / (0.5*(maxYdataAbs + minYdataAbs));

I am not happy with neither one.

Assume that you have the following data.

enter image description here

Only one or two sharp peaks. I am not happy with #1 or #2.

I am not sure anymore if the best way is to normalise to the range $[-1,1]$.

How should you normalise such a data?

Improve this question
$\endgroup$
1
  • 4
    $\begingroup$ Why aren't you satisfied with these approaches? When you're normalizing data, you're simply scaling it and/or adding offset. Normalizing will not change how your data looks. What exactly are you trying to achieve? $\endgroup$
    – Phonon
    Commented Nov 9, 2013 at 7:08

1 Answer 1

Reset to default
2
$\begingroup$

You could normalize the average amplitude, i.e.

YData = YData / mean(abs(YData));

Or you could normalize the signal power to one, i.e.

YData = YData / sqrt(mean(abs(YData).^2));

If just the peaks are bothering you, you could use dynamic range compression, but that would introduce nonlinear distortions. As Phonon hinted, please tell us why you are not happy with the results and why you want to normalize your data in the first place. That would allow for a more precise answer.

Improve this answer
$\endgroup$
1
  • $\begingroup$ You understand this topic better than me. Those nonlinear distortions are bothering me. I want to minimise distortions well, but I want to measure their degrees. I do not believe that to minimise one type of distortions makes the data most reliable. In short, I need to minimise the interferences in time-frequency analysis, which requires exact understanding which kind of distortions specific kind of normalisation is causing. So please point me to some direction if you know more about the topic. $\endgroup$
    – Léo Léopold Hertz 준영
    Commented Nov 9, 2013 at 21:52

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.