Until recently, signal processing involving successive time delays was implemented either by use of physical analog delay elements, such as acoustic or electric transmission-line delay elements, or by conversion to digital format for processing, then reconverting the processed results to an analog output. Wholly analog systems are cumbersome, expensive, and sensitive to environmental factors. On the other hand, conversion to wholly digital format is often unnecessary and penalizes the system in terms of cost, complexity, speed, and power. A very attractive alternative is discrete-time processing of analog samples. Time is quantized, but relative amplitudes are preserved. Delay is accomplished by transferring samples from cell to cell in shift-register fashion, while preserving relative amplitudes. Discrete-time systems combine many of the best features of both digital and analog systems: speed as well as the freedom from quantization effects of analog systems are combined with the time percussion and flexibility of digital systems.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.