2 Signal processing

2.5 Analog - digital

Processing of signals can be done in two ways: analog or digital. Real-world signals are analog, so processing them with analog electronics is the most obvious - and before digital signal processing (DSP) was possible - the only way. When DSP became available, possibilities grew immensely and they're still growing, because the limit is determined only by the computational speed of the applied digital processor(s), and computing speed doubles every 18 months nowadays. Digital processing requires highly complex designs, and still some analog circuitry, because the real-world signals have to be converted to digital signals (sampling or ADC, Analog-Digital-Conversion) first, then processed by the actual processor(s), and afterwards converted back to analog (DAC, Digital-Analog-conversion).

ADC, in short, is measuring the signal level many times per second, too fast to be descerned by the human ear, and describing each measurement with a number. The resulting sequence of numbers is often called "stream". DSP, in short, is applying complex math to the stream. DAC, in short, is translating the values in the stream back to signal levels.

Some charateristics of each type are in the table below:

Analog Digital
Extremely simple designs are possible to achieve the end result Designs require a minimum level of (high) complexity
More complex designs generally require more electronics More complex designs hardly need more electronics - the opposite seems to be true: as science progresses, systems become more compact, but with improved capability
Susceptible to noise and interference Once digital, much less noise is added, but strong interference may lead to hard failure
Capabilities are limited Only the designer's imagination and the current computing speed limit capabilities. There's a trade-off between sound quality and capability

© Joris van den Heuvel 2001-2009