In addition to the noise already mentioned in previous
sections, another effect that appears in the digital channel is distortion.
Given that the channel has a bandwidth, that is, it lets a limited number of frequency components of a
signal pass through it unmodified, any pulse that passes through it will be broadened. This
broadening is not noticable if the bandwidth of the channel is much greater than the bandwidth
of the pulse spectrum. If both bandwidths are comparable, however, the pulse at the output of the
channel will be noticably broadened. This means that when looking at a series of consecutive
pulses, each of them will overlap with the ones next to it, and this effect provokes errors (a
"0" may be interpreted as a "1"). In short, this will increase the probability
of errors in the bit and lower the transmission quality.
This effect is a type of distortion known as Inter-Symbol Interference (ISI) and was studied
by Harry Nyquist, who demonstrated that the bandwidth
B needed to avoid it, transmitting a specific form of signal in an ideal channel, when the
symbol transmission rate is Rs symbols per second, is as described in the following expression:
[B]: Hz
[Rs]: symbols /s
A symbol is a waveform belonging to a limited set of waveforms that is used as part of an
alphabet for transmitting digital information. An example of a symbol would be the combination
of zeros and ones that represents (symbolizes) each of the characters on a computer keyboard.
In another case, a symbol may represent each of the levels in which the amplitude of any
analog signal can be broken down into for its digitalization.
