I have a process which is most probably red-noise and my goal is to know if the peaks or pits are "real" or just due to random fluctuations.
I would like to know if there is exists a test of statistical significance for the peaks of a power-spectral density, obtained by welch's method for example. Or even better : a way to have error bars.
I have search the litterature for the past 4 hours and found very few ressources, that even contradict themselves sometimes...
Any help will be really appreciated :)