Digital Signal Processing Reference
In-Depth Information
Signal
Signal
Analog
Memory
(a)
Signal
Numbers
Numbers
Signal
Digital
Memory
ADC
DAC
Sample at Rate Fs
Output Samples
at Rate Fs
(b)
Figure 1.1:
(a) Conventional analog recording and playback system; (b) A digital recording and playback
system.
1. Analog hardware, such as amplifiers, filters, comparators, etc., is very susceptible to noise and
deterioration through aging. Digital hardware works with only two signal levels rather than an infinite
number, and hence has a high signal to noise ratio. As a result, there is little if any gradual deterioration
of performance with age (although as with all things, digital hardware can suddenly and totally fail),
and copies of signal files are generally perfect, absent component failure, media degeneration or damage,
etc. This is not true with analog hardware and recording techniques, in which every copy introduces
significant amounts of additional noise and distortion.
2. Analog hardware, for the most part, must be built for each processing function needed. With
digital processing, each additional function needed can be implemented with a software module, using
the same piece of hardware, a digital computer. The computing power available to the average person
has increased enormously in recent years, as evidenced by the incredible variety of inexpensive, high
quality devices and techniques available. Hundreds of millions or billions of operations per second can
be performed on a signal using digital hardware at reasonable expense; no reasonably-priced alternative
exists using analog hardware and processing.
3. Analog signal storage is typically redundant, since wave-related signals (audio, video, etc.) are
themselves typically redundant. For example, by taking into account this redundancy as well as the
physiological limitations of human hearing, storage needs for audio signals can be reduced up to about
95%, using digitally-based compression techniques, such as MP3, AC3, AAC, etc.
4. Digital processing makes possible highly efficient security and error-correction coding. Using
digital coding, it is possible, for example, for many signals to be transmitted at very low power and to
