Systems and Filters

Within the `signal phonetic tape' model, speech signal processors slot into the `AS', the `acoustic signal' point, transform the analogue (continuous) signal into a digital (discrete) signal, process this, and possibly transform it back into an analogue signal for further transmission to a receiver.

1. The source of a signal is modelled as a system with an acoustic output. A system which models an operation on a signal is a filter, which can be thought of as a black box with one or more inputs and one or more outputs.
2. An analog system or continuous-time system processes continuous-time signals, while a digital system or discrete-time system processes discrete-time signals.
3. An analog-to-digital converter, ADC, samples a continuous-time signal, generally at regular intervals, and converts it to a sequence of discrete-time values. The sampling frequency or sampling rate must be at least twice as high as the highest frequency to be measured, or an effect known as the `aliasing effect' occurs: a phantom signal appears in the output. For example, if the frequency to be measured is an exact multiple of half a wavelength, then the signal will appear to have the same invariant amplitude.
   Conversely, a digital-to-analog converter, DAC, converts a digital signal to an analog signal.
4. A digital signal processor, DSP, is a dedicated computer, often on a single chip, for performing programmable operations and transformations on digital signals.
5. A digital signal processing system has the structure shown in the Figure. The analogue inputs and outputs are filtered through low-pass filters: the input filter is the anti-aliasing filter, used to cut out signals which are too high in frequency to be sampled properly, and the output filter cuts out the sampling frequency from the system output.

    
   Figure 4: Block diagramme of a digital signal processing system.