Obtaining the Frequency Content of a Sampled Sound

[jvalentino]

This is a tutorial I wrote on how to obtain the frequency content of a sampled sound using the discrete fourier transform in Java.

This is the Link to the Tutorial

This is a good starting point for those who are interested in sound processing and analyzation, and have a little background in math and computer science.

[reith]

Is the bottom diagram the sectrum of the signal shown? What are the units? I gather the x axis (1) is time but 0.2 (etc) what?

Assuming the frequency is constant over a time interval, if all you want is to extract the frequency, isn't it easier to use a frequency counter.

[Ron Ofir]

I think that's the wave function.

[jvalentino]

Sorry, I just pulled the images from the link without labeling them. The top image is the waveform, which is pressure (y) over time in seconds (x). The bottom image is the resulting frequency spectrum of the above waveform as amplitude (y) over frequency (x).

It uses audio data stored in the WAV format, meaning any recorded sound. The intention is for the use of complex sounds.

A frequency counter is a hardware solution involving a circuit, not a software solution.

[reith]

Complex sounds? Like complex waveforms at a fixed frequency or varying like audio?

[jvalentino]

Yes complex waveforms, meaning that they contain more than one frequency either through harmonics or from other sound producing sources.

It will work for both a sound sustaining the same pitch over a given amount of time or for a sound changing pitch. The thing about Fourier transforms though is that it takes data from time domain into the frequency domain by converting the signal information to a magnitude and phase component of each frequency.

This means that time (as seen in the first graph) on the x axis becomes frequency (as seen in the second graph). Without getting into the math, the sampling time in combination with some other factors what how frequencies can be detected. Furthermore any frequency that occurs during the sample time appears in the resulting spectrum.