Anyone been watching soccer lately? If so, you may have noticed a buzz in the background, created by many people playing the vuvuzela. This contribution on the File Exchange by Choqueuse Vincent might help you filter the sound track! Vincent uses the spectral subtraction method to achieve some nice results.
Contents
Sample Data
Here's a spectral plot including contributions from many vuvuzelas.
What Kind of Noise Does Your Data Have?
When I was studying Earth's magnetic field, we always joked that one person's noise was someone else's signal - because the main contributions to the magnetic field are crustal contributions, deep Earth contributions, and those external to the planet. I was at various times trying to model or understand contributions from Earth's interior (sometimes crustal, sometimes core) while others were studying the external sources. Because we could apply at least a simple model to the sources we weren't studying, we were each able to model the sources of interest.
Do you have the luxury of understanding enough about the sources of noise in your data? Let me know here.
Get
the MATLAB code
Published with MATLAB® 7.10
I like to know pretty much!
That’s where my focus of research is.
Please let me know?
Regards,
I study passive sonar. All we have is noise!
Loren, you’ve committed the cardinal sin of not labelling or dimensioning your axes :-) What am I looking at?
Mark-
That’s the plot from the File Exchange submission so you’ll have to ask the author, Choqueuse Vincent.
Colin-
True, there are some domains without an explicit “signal”!
Abiyu-
You might contact the author of the file on the File Exchange to get more information from him.
–loren
@Mark:
It seems to be a new trend here on the blogs. All the cool kids are doing it :P
Is there a way to de-noise spectrogram?
Michael-
Check out Signal Processing Toolbox for all kinds of filtering techniques.
–Loren
I’ve now added the spectrogram axes :)
Choqueuse-
Thanks! People will appreciate that. And here’s a link to the new thumbnail, including the new annotations.
–Loren
Thanks Loren !