I have the following data values
freq=[10 30 100 300 1000 3000 10000]; %frequency (Hz)
ew = [2 1.4 1.4 1.1 1.1 1.1 1.1]*1e-9; %amplitude %nV/sqrt(Hz)
Based on this data values and assuming a modal of 1/frequency (pink noise). I wish to extrapolate the data at earlier frequencies 0.01-10Hz. Other information I have is the data from 0.01-10Hz has a peak-peak amplitude of 80 nV. Therefore the integral between 0.01-10Hz is a constraint. Is this possible to do?
So I will be extrapolating and fitting between freq=logspace(-2,1e4,1000)
Currently I'm having a look at SLM - Shape Language Modeling and optimization toolbox. The key is to ensure the constraint of 80nV p-p is closely satisfied. At frequency=0 needs to be ignored since this will blow up.