expand

Average

Wave
Spectrum

    

This is an example of a two-point average filter. Filters are used to shape noise or other sounds, so that they become louder at some frequencies and softer at others. The two-point average filter replaces every sample of the wave with the average of the sample and the one before it. Mathematically, this says output[t] = 0.5 * (input[t] + input[t-1]).

Averaging smooths out the wave a little bit, which eliminates the highest frequencies, but doesn’t affect the low frequencies. The audio spectrum isn’t flat like for white noise, but instead goes down to zero at the high end.

So a two-point average (also sometimes called a ‘one-zero filter’) is the simplest possible ‘low pass filter’. However, it doesn’t do as good a job of filtering out high frequencies as a standard low pass filter.