MOVING AVERAGE
FILTERS
2
Moving Average Filter
The moving average filter operates by averaging a number of
points from the input signal to produce each point in the output
signal.
In equation form , this is written:
In this equation, x[ ] is the input signal, y[ ] is the output signal, and
M is the number of points used in the moving average.
This equation only uses points on one side of the output sample
being calculated.
3
Moving Average Filter
For example, in a 5 point moving average filter, point 80 in the
output signal is given by:
As an alternative, the group of points from the input signal can
be chosen symmetrically around the output point:
This corresponds to changing the summation from: j= 0 to M-1,
to: 𝑗 = −(𝑀 − 1)/2 𝑡𝑜 (𝑀 − 1)/2 .
Symmetrical averaging requires that M be an odd number.
4
Moving Average Filter
The moving average filter is a convolution using a very simple
filter kernel.
For example, a 5 point filter has the filter kernel: …, 0, 0, 1/5, 1/5,
1/5, 1/5, 1/5, 0, 0, …
The moving average filter is a convolution of the input signal
with a rectangular pulse having an area of one.
5
Noise Reduction vs. Step Response
Moving average filter is very good for many applications.
It is optimal for a common problem, reducing random white
noise while keeping the sharpest step response.
Example of using a moving
average filter.
In (a), a rectangular pulse
is buried in random noise.
6
Noise Reduction vs. Step Response
In (b) and (c), the smoothing action of the moving average filter
decreases the amplitude of the random noise (good), but also reduces the
sharpness of the edges (bad). Of all the possible linear filters that could be
used, the moving average produces the lowest noise for a given edge
sharpness. The amount of noise reduction is equal to the square-root of the
number of points in the average. For example, a 100 point moving average
filter reduces the noise by a factor of 10.
7
Frequency Response of MA Filter
Frequency response of an M point
moving average filter is mathematically
described by the Fourier transform of
the rectangular pulse:
The frequency, f, runs between 0 and 0.5.
For f=0, 𝐻[𝑓] = 1
The roll-off is very slow and the stopband attenuation is bad. Clearly, the
MA filter cannot separate one band of frequencies from another. Good
performance in the time domain results in poor performance in the
frequency domain, and vice versa. In short, the MA is an exceptionally good
smoothing filter (time domain), but an exceptionally bad low-pass filter
(frequency domain).
8
Relatives of the Moving Average Filter
Filter designers would only have to deal with time domain or
frequency domain encoded information, but never a mixture of
the two in the same signal. Unfortunately, there are some
applications where both domains are simultaneously important.
Television signals (video information)
Electromagnetic interference
Relatives of the moving average filter have better frequency
domain performance, and can be useful in these mixed domain
applications.
9
Multiple-pass moving average filter
Multiple-pass moving average filters
involve passing the input signal
through a moving average filter two
or more times.
Figure shows the overall filter kernel
resulting from one, two and four
passes.
Two passes are equivalent to using a
triangular filter kernel (a rectangular
filter kernel convolved with itself).
After four or more passes, the
equivalent filter kernel looks like a
Gaussian (recall the Central Limit
Theorem).
10
Multiple-pass moving average filter
As shown in figure,
multiple passes produce an "s"
shaped step response, as
compared to the
straight line of the single pass.
Multiple-pass moving average filter
11
The frequency responses in (c) and
(d) are given by H(f) multiplied by
itself for each pass.
Each time domain convolution results
in a multiplication of the frequency
spectra.
Gaussian Filter
12
When a pure Gaussian is used as a
filter kernel, the frequency response is
also a Gaussian
The Gaussian is important because it is
the impulse response of many natural
and manmade systems.
For example, a brief pulse of light
entering a long fiber optic transmission
line will exit as a Gaussian pulse, due
to the different paths taken by the
photons within the fiber.
The Gaussian filter kernel is also used
extensively in image processing
because it has unique properties that
allow fast two-dimensional
convolutions
13
Blackman window
The second frequency
response in Figure
corresponds to using a
Blackman window as a filter
kernel. (The term window has
no meaning here; it is simply
part of the accepted name
of this curve).
The exact shape of the
Blackman window looks
much like a Gaussian.
14
Сomparative analysis of MA filters
How are these relatives of the moving average filter better
than the moving average filter itself?
1) these filters have better stopband attenuation than the moving
average filter.
2) the filter kernels have a smaller amplitude near the ends.
Recall that each point in the output signal is a weighted sum
of a group of samples from the input. If the filter kernel tapers,
samples in the input signal that are farther away are given less
weight than those close by.
3) the step responses are smooth curves, rather than the sharp
straight line of the moving average.
15
Сomparative analysis of MA filters
The moving average filter and its relatives are
all about the same at reducing random noise.
The ambiguity lies in how the risetime of the
step response is measured. If the risetime is
measured from 0% to 100% of the step, the
moving average filter is the best you can do,
as previously shown.
In comparison, measuring the risetime from
10% to 90% makes the Blackman window
better than the moving average filter.
Consider these filters equal in this parameter.
16
Сomparative analysis of MA filters
The biggest difference in these filters is execution speed.
Using a recursive algorithm, the moving average filter will run
like lightning in your computer. In fact, it is the fastest digital
filter available.
Multiple passes of the moving average will be correspondingly
slower, but still very quick.
In comparison, the Gaussian and Blackman filters are slow,
because they must use convolution. Think a factor of ten times
the number of points in the filter kernel (based on multiplication
being about 10 times slower than addition). For example,
expect a 100 point Gaussian to be 1000 times slower than a
moving average using recursion.
17
Recursive Implementation
A tremendous advantage of the moving average filter is that it
can be implemented with an algorithm that is very fast.
Imagine passing an input signal, x [ ], through a M=7 point
moving average filter to form an output signal, y [ ].
Now look at how two adjacent output points, y [50] and y [51],
are calculated:
If y [50] has already been calculated, the most efficient way to
calculate y [51] is:
18
Recursive Implementation
After the first point is calculated in y [ ], all of the other points
can be found with only a single addition and subtraction per
point. This can be expressed in the equation:
Notice that this equation use two sources of data to calculate
each point in the output: points from the input and previously
calculated points from the output.
This is called a recursive equation, meaning that the result of
one calculation is used in future calculations.
Recursive Implementation
19
This algorithm is faster than other digital filters for several reasons.
1) There are only two computations per point, regardless of the
length of the filter kernel.
2) Addition and subtraction are the only math operations needed,
while most digital filters require time-consuming multiplication.
3) The indexing scheme is very simple. Each index is found by
adding or subtracting integer constants that can be calculated
before the filtering starts (i.e., p and q).
4) The entire algorithm can be carried out with integer
representation. Depending on the hardware used, integers can
be more than an order of magnitude faster than floating point.
5) Integer representation works better than floating point with this
algorithm, in addition to being faster. There is no round-off error in
the arithmetic.
