MATLAB Answers

Running Average of a Massive data set

2 views (last 30 days)
So I have a data file containing 5 seconds of sound data, where I want to smooth the data with a running average function and then use the Fast Fourier Transform function in MATLAB in order to pinpoint the amplitude spikes that should be a tuning fork and an out of tune instrument. The data has 20,000 popints, however, and the methods I've tried so far accidentially erase the information I want. Is there a way to make a running average function manually for that many data points?


Sign in to comment.

Accepted Answer

Image Analyst
Image Analyst on 10 Dec 2019
Well 20,000 points is far from massive. Maybe if it were 10,000 times that big. 20k is actually pretty small. Anyway you can use the conv() function.
windowWidth = 101; % an odd number.
kernel = ones(1, windowWidth) / windowWidth;
smoothedSignal = conv(signal, kernel, 'same');
You can also try the movmean() function if you have a recent version of MATLAB.

  1 Comment

Duncan Cross
Duncan Cross on 10 Dec 2019
It's the biggest one I've had to work with, so it seems massive to me. But your code does smooth it pretty well, thank you!

Sign in to comment.

More Answers (0)

Sign in to answer this question.