Hi

I believe STFT simply is ordinary FFT but where the window slides with data instead of averages data, i.e. if you have a signal X and makes a first FFT of samles 1 to N.

The ordinary FFT would continue by averaging working on samples  N+1 to 2N and so on.

STFT does instead jump to, say, samples 10 to N+9 and make a new FFT, then move on to samples 20 to N+19 and so on. The result would be a colorplot or waterfall diagram that shows how the signal evolves with time.

https://www.math.ucdavis.edu/~strohmer/research/gabor/gaborintro/node3.html

http://www.originlab.com/index.aspx?go=Products/Origin/DataAnalysis/SignalProcessing/STFT

If you have access to matlab, take a look here

http://www.mathworks.com/matlabcentral/fileexchange/45197-short-time-fourier-transformation--stft--with-matlab-implementation

https://www.youtube.com/watch?v=NA0TwPsECUQ

You can download also free apps for your smartphone that uses the microphone and shows spectrograms, which is kind of fun (if you are so inclined).

Sincerely

Claes

You could adopt the methodology from Renaudin et al, Step length estimation using handheld inertial sensors, Sensors (2012) or Mitchell et al, Automatically Detecting Asymmetric Running using Time and Frequency Domain Features, IEEE 2015

Thank you very much for your answer, I'll use some of the document that you has share with me. I'm not mathematician, but I'll try to apply it with matlab.

I have some other questions: If I would like to apply the STFT in a running support phase (no more than 350 ms), do you know what kind of window I have to use?, what's the best samping rate to apply the STFT?

Regards and sorry if the questions are so elemental!

Alberto

Should help get you started - best thing to do is trial and error - give it a go!!

https://plot.ly/matlab/stft/

http://uk.mathworks.com/matlabcentral/fileexchange/45197-short-time-fourier-transformation--stft--with-matlab-implementation

http://uk.mathworks.com/help/signal/ref/spectrogram.html

http://www.mathworks.com/matlabcentral/fileexchange/45197-short-time-fourier-transformation-stft-with-matlab-implementation/content/example.m

https://www.youtube.com/watch?v=3G4sl_oaZT4

Remember - there are lots of Matlab tutorials on line (like the YouTube video above) and "MatlabCentral" is an extended resource for source code.