Time and Frequency analysis

Hello,

I have data of measured signal in time. I want to know the frequencies of the components of my signal (Large scale fading+Small scale fading+Noise). For this reason, I have represented the FFT of my signal. I would like to know the time instant where the peaks highlited in yellow happened. Does any one have an idea how to do it?

Another question, How can I extract each component of the original signal from the FFT representation?

Thank you in advance.

Answer 1

[Star Strider]

The peaks in the fft result can occur anywhere in the signal. To do a time-frequency analysis, use the spectrogram function.

[yusra Ch]

You mean like this? Do you know how to represent the frequency axis in Hertz?

[Star Strider]

Yes. Experiment with the different parameters and name-value pair arguments to get different results.

To get the frequencies in Hz, specify the sampling frequency, fs. See: Spectrogram and Instantaneous Frequency.

[yusra Ch]

I have specified these parameters:

Fm=1000;
Freq=0:Fm/length(potencia(L,:)):Fm/2;
spectrogram(power,[],[],Freq,Fm,'yaxis');

I know that the spectrogram explains how the signal strength is distributed in every frequency found in the signal. However, I don't know how can I extract information about the large scale fading, small scale fading and noise.

[Star Strider]

That is likely not possible to extract from a spectrogram.

This sounds like a communications problem rather than the signal processing problem I first thought it was.