Digital Signal Processing Reference
In-Depth Information
Time domain
Spectrum
Freq. domain
Nois e
Integration
5,00
2,50
0,00
-2,50
-5,00
5,00
2,50
0,00
-2,50
-5,00
0,50
0,25
0,00
-0,25
-0,50
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
850
900
950
ms
Illustration 136:
Forming mean value by integration
In this case a noise signal is integrated. The result is practically zero with small fluctuations (bottom
higher resolution. The process of integration adds the positive and negative „areas“ together. Because
everything is random in the case of a noise signal an equal distribution of positive and negative „areas“
must be present. In the average they must all be zero.
Accordingly, integration can be used to form the mean value, an extremely important aspect for measuring
technology. In addition, forming a mean value apparently eliminates noise better than a normal filter
could.
It is striking how quickly the mean value zero is reached in the case of noise. As noise consists of a
„stochastic sequence of individual pulses“, i.e. practically of a sequence of weighted
pulses, this is easy
to understand. Positive and negative „areas“ which have an important effect only arise when a state is
maintained „for not inconsiderable periods of time“.
δ−
In the examination of the integration process as a generator of mean values Illustration
136 arrives at the following result:
•
As a result it is clear that noise has the mean value of zero. The integra-
tion of a noise signal practically gives this value, i.e. the mean value
can be determined by means of integration.
•
Apparently, noise elements can be removed from a signal by means of
integration, better than this is possible with normal filters.
