FIR filters use a running weighted average to form the output from the
input. In this demo we use a 5-point averager which computes the
output \(y[n] \) from the input \(x[n]\) via the difference
equation:
$$y[n] = \frac{x[n] + x[n-1] + x[n-2] + x[n-3] + x[n-4]}5$$

Below is the time-domain plot and the spectrum of the input signal. Do they make sense?

You can listen to the input signal that will be passed through the 5-point averager.

Below is the time-domain plot and the spectrum of the input signal. Do they make sense?

Here is the *frequency response* of the 5-point averager.

It lets the lower frequencies pass through and attenuates (makes
softer) the higher frequencies. Some frequencies are completely
removed. Which ones are they?

Listen to the signal after filtering.

Compare to the input signal. Why do they
differ? Can you hear where the frequencies are missing?

Below is the spectrum of the output signal (red and dark red indicate
the largest values, while green and blue are small). Can you see which
frequencies have been removed?