Inter-symbol Interference

Inter-symbol interference (ISI) is an unavoidable consequence of both

wired and wireless communications systems.

It is noticed that the pulses after transmission tend to get elongated and
smeared into each other when they reach the receiver. A short pulse is
received as a much-smeared version of the same thing. The problem is
related to the properties of the medium used and the distance of signal
travel.
(a) What was sent, (b) What was received.
Suppose, we wish to send a data sequence 1,0,1,1,0. This sequence is in
the form of square pulses. Square pulses are hard to create in practice
and also require far too much bandwidth. So we shape them as shown in
the dotted line. The shaped version looks essentially like a square pulse
and we can quickly tell what was sent even visually. Advantage of
shaping at this point is that it reduces bandwidth requirements and can
actually be created in practice.
Sequence 101101 to be sent, the dashed line is the shape that is actually sent
Next figure shows each symbol as it is received. We can see that the
transmission medium creates a tail of energy that lasts much longer than
intended. The energy from symbols 1 and 2 goes all the way into symbol
3. Each symbol interferes with one or more of the subsequent symbols.
The circled areas show areas of large interference.
Each symbol is spread by the medium
Sequence 101101 to be sent, the dashed line is the shape that is actually sent

Each symbol is spread by the medium

This figure shows the actual signal seen by the receiver. It is the sum of
all these distorted symbols. Compared to the dashed line that was the
transmitted signal, the received signal looks quite indistinct. The receiver
does not actually see the signal, it sees only the little dots, the value of the
amplitude at the timing instant. Notice that for symbol 3, this value is
approximately half of the transmitted value, which makes this particular
symbol more susceptible to noise and incorrect interpretation and this
phenomena is the result of this symbol delay and smearing.
Sequence 101101 to be sent, the dashed line is the shape that is actually sent

Each symbol is spread by the medium

Received signal vs. the transmitted signal

Due to the spreading and smearing of symbols, the energy from one
symbol effects the next ones in such a way that the received signal has a
higher probability of being interpreted incorrectly. This is called Inter
Symbol Interference or ISI. ISI can be caused by many different reasons.
It can be caused by filtering effects from hardware or frequency selective
fading, from non-linearities in the channel, from multipath propagation,
etc. Very few systems are immune from it and it is nearly always present
in wireless communications.
Communication system designs for both wired and wireless nearly
always need to incorporate some way of controlling the effects of ISI.
The main problem is that energy, which we wish to confine to one
symbol, leaks into others. So one of the simplest things we can do to
reduce ISI is to just slow down the signal. Transmit the next pulse of
information only after allowing the received signal to damp down. The
time it takes for the signal to die down is called delay spread, whereas the
original time of the pulse is called the symbol time. If delay spread is less
than or equal to the symbol time then no ISI will result, otherwise there
will be.
Slowing down the data rate is an easy but an unacceptable solution. One
way to counter ISI is pulse shaping. When the timing pulse samples the
signal to determine the value of the signal at that instant, it does not take
into account what the signal looked like before or after it. So if there was
some way we could keep the symbols from interfering in such a way that
they do not affect the amplitude at the sampling instant, we can counter
ISI successfully. There are other signal processing techniques to mitigate
the effect of ISI
Noise
 As discussed in the analog voice channel, here also thermal noise, impulse
noise, crosstalk etc. affect system design.
 These impairments need only be considered on a per-repeater-section basis
because noise does not accumulate due to the regenerative process carried out
at repeaters and nodes.
 But bit errors do accumulate, and these impairments among others create
these errors.
 One way of limiting error accumulation is to specify a stringent BER for each
repeater section. This value is required to ensure the correct operation of
supervisory signaling.
Quantization Noise
It is the primary source of noise in a PCM system. The signal-to-noise
ratio (or more precisely signal-to-quantization noise ratio), SQR, in a
system using linear quantization is relatively easy to determine and is
given by the following expression, where n is the number of bits /
sample, and a sinusoidal signal is assumed:
SQR = 6n + 1.8 dB
The SQR is a linear function of signal level. A smaller signal
level than the assumed above naturally results in a smaller SQR value.
 Aj+a/2

th
Aj j level

Aj-a/2
a

Aj+

Aj- receiver output at quantized voltage, Aj+- instantaneous voltage of the

signal, - equally likely error voltage
Mean squared error
1 a/2 2 a2
E ( ) 
2
  d 
a a / 2 12 the average quantizing noise power
a2 a
 
2
 Pq , RMS  2
12 3
SQR for sinusoidal modulation:

Am 2
amplitude of sine wave Am , average signal power
2
2 Am
peak-to-peak excursion is 2 Am a ,
L
Am 2
2
Am 2  ( SQR)  22 
3 L
No of levels is L  Pq  2 out
Am 2
3L
2
3L
 ( SQR ) out dB  1.8  20 log L  1.8  6n dB for L  2n
 When digitizing a voice channel, however, nonlinear quantization is
employed (so-called companding = compression before A/D conversion +
expanding after D/A conversion). This means that quantization intervals at
low signal levels are closely spaced and are correspondingly very widely
spaced at high signal levels.

Companded PCM with compression and expansion

 Typical compression characteristic for speech encoding
One of the main reasons of utilizing the principle of companding is to
obtain a reasonably constant SQR value with respect to signal level. As
indicated in the figure, the use of companding together with n = 8 bits / sample
provides sufficiently large SQR for voice digitization purposes.
Signal-to-(quantizing) noise ratio, SQR versus input signal power:
(a) 8-bit PCM using companding (m-law with m = 255),
(b) linear 8-bit PCM and
(c) linear delta modulation.