## Finding the spectrums of three segments of a signal without using the 'fiindpeaks' function?

### Hannah Oduntan (view profile)

on 9 Jan 2019
Latest activity Commented on by TADA

on 11 Jan 2019
How can I find the spectrums of the segment of the signal before the major peak, after the major peak and othe spectrum of the major peak all separately? Doing all this without using the findpeaks function. The signal link has been attached.

### Tags

on 9 Jan 2019

on 9 Jan 2019

There can be quite a few approaches to do this
A simple yet effective method would be to analyze the slope, you can start by finding the peak location by finding the maximum signal value (granted this is a single peak signal)
then you can detect the edges of the peak by tracing changes to the slope from positive to negative
If the signal may contain more than one peak, you can use the maximal changes in the slope to locate the peak locations as well.
you can use that as a starting point:
x = 1:length(Signal);
y = Signal;
peakIdx = find(y == max(y));
dy = diff(y);
peakStart = find(dy(1:peakIdx-1) < 0, 1, 'last');
peakEnd = find(dy(peakIdx:end) > 0, 1, 'first') + peakIdx;
plot(x(1:peakStart), y(1:peakStart), '-b',...
x(peakStart:peakEnd), y(peakStart:peakEnd), '-g', ...
x(peakEnd:end), y(peakEnd:end), '-r');

Show 1 older comment

on 10 Jan 2019
Do you mean frequency spectrum?
Hannah Oduntan

### Hannah Oduntan (view profile)

on 11 Jan 2019
Yes, the frequency spectrum.

on 11 Jan 2019

### Sean de Wolski (view profile)

on 9 Jan 2019
Edited by Sean de Wolski

on 11 Jan 2019

doc pwelch
doc islocalmax
doc islocalmin