System Representation

Speech Production

We have tried to represent the speech production using the source filter theory. According to it larynx acts as source and the vocal tract as a filter and the combination of the two gives the speech.

Thus we can represent the speech production as a combination of three systems.

· Source system:-
The input is air given by the lungs.
The output is air pressure pulses.

· Filter system(an LSI system):-
The input is air pressure generated from the source.
The output is speech produced as air pressure fluctuations.

· System describing the effects of changes in vocal tract
The input is the constriction and the length of the tract.
The output is the frequency response of the filter.
Once the frequency response of the filter is decided by the constriction system the filter behaves as a LSI system. We could consider the filter system as a convolution of the two inputs the frequency response and the input obtained from source. In this presentation it is assumed, when talking about the filter system, that its frequency response is fixed and not considering it as an input.

Hearing System

We tried to represent Ear as a Bank of Band Pass Filters. According to it, the cochlea is represented as bank of Band Pass Filters.

The ear can be represented as a cascade of two systems:
a) The one formed by the outer and middle ear:-
The input is the sound wave
The output is also sound wave.
This system just amplifies the input.
b) The inner ear:-
The input is the output of the above system.
The output is neural signals.
This system acts as a bank of band pass filters.
These are a non linear filter which distills the various frequencies and send them to the brain.