Wireless Pi

DSP, Wireless Communications, SDR and More

https://wirelesspi.com

Minimum Shift Keying (MSK) – A Tutorial

By Qasim Chaudhari

Minimum Shift Keying (MSK) is one of the most spectrally efficient modulation schemes available. Due to its
constant envelope, it is resilient to non-linear distortion and was therefore chosen as the modulation
technique for the GSM cell phone standard.

MSK is a special case of Continuous-Phase Frequency Shift Keying (CPFSK) which is a special case of a
general class of modulation schemes known as Continuous-Phase Modulation (CPM). It is worth noting that
CPM (and hence CPFSK) is a non-linear modulation and hence by extension MSK is a non-linear modulation
as well. Nevertheless, it can also be cast as a linear modulation scheme, namely Offset Quadrature Phase
Shift Keying (OQPSK), which is a special case of Phase Shift Keying (PSK). As a borderline case, these
relationships are illustrated in Figure below.
 Figure 1: MSK as a special case of both non-linear and linear modulation schemes

At this point, you would be thinking about the following question: How can a modulation be both non-linear
and linear? As we see later in this tutorial, originally MSK is a non-linear modulation but a certain
depiction of its actual digital symbols known as pseudo-symbols turns it into an OQPSK representation.

Modulation is a simple topic to understand but owing to the above description, MSK can sometimes be an
intimidating concept. Here, our purpose is to present it in an uncomplicated manner by building it through the
fundamentals.

The starting point is one of the simplest digital modulations possible: FSK.

Binary Frequency Shift Keying (BFSK)

In Frequency Shift Keying (FSK), digital information is transmitted by changing the frequency of a carrier
signal. Naturally, Binary FSK (BFSK) is the simplest form of FSK where the two bits 0 and 1 correspond to
two distinct carrier frequencies F0 and F1 to be sent over the air. The bits can be translated into symbols
through the relations

0 → −1

1 → +1

This enables us to write the frequencies Fi with i ϵ {0, 1} as

i+1
F i = F c + (−1) ⋅ ΔF = F c ± ΔF

where Fc is the nominal carrier frequency and ΔF is the peak frequency deviation from this carrier
frequency. Consequently,

s(t) = A cos 2πF i t = A cos [2π{F c ± ΔF }t] —− Eq (1)

where
 0 ≤ t ≤ Tb

Figure 2 below displays a BFSK waveform for a random stream of data at a rate of Rb = 1/T b . Note that we
are not distinguishing between a bit period and a symbol period because both are the same for a binary
modulation technique.

Figure 2: A Binary Frequency Shift Keying (BFSK) waveform

As is evident from Figure 2 above, the phase transitions at the boundaries of bit transitions are — in general
— discontinuous.

Minimum Frequency Spacing

Although any two distinct frequencies F0 and F1 can be used for communication purpose, it greatly helps in
receiver design if the two distinct signals are orthogonal to each other, i.e.,

Tb

∫ s1 (t)s0 (t) dt = 0

A question that arises at this stage is the following: how close can the two frequencies F0 and F1 be? Or in
other words, what is the smallest possible value of ΔF ? The reason for asking this question is spectral
efficiency. The closer the two frequencies, the more the number of channels available for other users in the
same spectrum.

For a BFSK case,

Tb

∫ cos 2πF 1 t cos 2πF 0 t dt = 0

or
 Tb Tb

∫ cos 2π(F 1 + F 0 )t dt + ∫ cos 2π(F 1 – F 0 )t dt = 0

0 0

Tb Tb
sin 2π(F 1 + F 0 )t sin 2π(F 1 − F 0 )t
∣ ∣
∣ + ∣ = 0
2π(F 1 + F 0 ) ∣ 2π(F 1 − F 0 ) ∣
0 0

sin 2π(F 1 + F 0 )T b sin 2π(F 1 − F 0 )T b

+ = 0
2π(F 1 + F 0 ) 2π(F 1 − F 0 )

Since F1 + F 0 = 2F c ≫ 1 while −1 ≤ sin x ≤ 1, the first term goes to zero and we can write

sin 2π(F 1 − F 0 )T b = 0

which is true for (remember sin kπ = 0 for integer k)

2π(F 1 − F 0 )T b = kπ

From here, the orthogonality condition can be concluded as

k
F1 − F0 =
2T b

This also yields the minimum frequency separation for k = 1.

1 Rb
F1 − F0 = =
2T b 2

for orthogonal signaling. Thus, the peak frequency deviation ΔF can be computed as

F1 − F0 1 Rb
ΔF = = =
2 4T b 4

MSK as Continuous-Phase FSK

From the above information, a BFSK signal with minimum tone spacing can be constructed by replacing ΔF
by Rb /4 in Eq (1) as

Rb
s(t) = A cos [2π{F c ± }t],
4
Rb
= A cos [2πF c t ± 2π t], 0 ≤ t ≤ Tb
4

This is a CP-BFSK signal with minimum tone spacing defined over a single bit interval 0 ≤ t ≤ Tb. There
are two more steps to construct an actual MSK waveform.

1. In a real communication system, the signal is constructed by transmitting a sequence of bits bn in

succession, where n is the bit index. As we have seen earlier, bits bn  ϵ {0, 1} are converted to symbols
a n .

a n  ϵ {−1, +1}
 Then, in the interval nT b ≤ t ≤ (n + 1)T b , the above signal can be written as

a n Rb
s(t) = A cos [2πF c t + 2π (t − nT b )], nT b ≤ t ≤ (n + 1)T b
4

where n is the bit index within a long bit stream and the second term indicates the underlying baseband
message.
2. Observe in the above equation that the phase continuity is not necessarily maintained from one symbol to
the next. To ensure phase continuity, we must add a phase component for each symbol as

a n Rb
s(t) = A cos [2πF c t + 2π (t − nT b ) + θn ]
4

nT b ≤ t ≤ (n + 1)T b —− Eq (2)

For this purpose, the phase at both sides of t = (n + 1)T b must be equal, as illustrated in Figure 3.

Figure 3: Phase on both sides of t = (n + 1)T b must be equal to maintain continuity

Thus, at the instant t = (n + 1)T b , the following equation must be satisfied.

a n Rb a n+1 R b
∣ ∣
2π (t − nT b ) + θn ∣ = 2π (t − (n + 1)T b ) + θn+1 ∣
4 ∣ 4 ∣
t=(n+1)Tb t=(n+1)Tb

which can be written as

π
θn+1 = θn + a n
2

Another way to write the above recursive relation is

π
θn = θn−1 + a n−1 —− Eq (3)
2

Assume that θ0 = 0. Then,

π
θ1 = a 0
2
 π π
θ2 = θ1 + a 1 = (a 0 + a 1 )
2 2

In general,

n−1
π
θn = ∑ ai —− Eq (4)
2
i=0

When the phase follows this rule during each bit/symbol interval, the phase continuity is ensured and the
resulting waveform is shown for an example sequence in Figure 4.

Figure 4: MSK as Continuous-Phase Binary FSK

Notice how Figure 4 is different than Figure 2 in phase continuity.Below, we plot θn as a function of time in
Figure 5 to see how it evolves. One can observe that it indeed changes values in steps of π/2 depending on
the last data bit.
 Figure 5: θn evolving with time

MSK as Offset QPSK

First, write Eq (2) assuming A = 1 as

a n Rb
s(t) = cos [2πF c t + 2π (t − nT b ) + θn ]
4
a n Rb π
= cos [2πF c t + 2π t– na n + θn ] —− Eq (5)
4 2

Θn

Using Eq (3), we can further manipulate Θn as

π
Θn = θn – na n
2
π π
= θn−1 + a n−1 – na n
2 2
π π
= θn−1 + a n−1 ( + 1 − n + n) – na n
2 2
π π
= θn−1 – (n − 1)a n−1 + n (a n−1 – a n )
2 2
π
= Θn−1 + n (a n−1 – a n ) —− Eq (6)
2
Notice that Θn = Θn−1 when

a n = a n−1 because the second term in Eq (6) becomes zero.

n is even because the second term in Eq (6) is a multiple of 2π (remember that a n−1 − an is ±2 when
not zero).

We conclude that Θn can only change when a n ≠ a n−1 and n is odd. In that case, it will always change by
an odd multiple of ±π. In summary, considering modulo-2π operations,

Θn−1 ± π a n ≠ a n−1 , n odd

Θn = { —− Eq (7)
Θn−1 otherwise

Next, we use the following identities

cos(α ± β) = cos α cos β ∓ sin α sin β

sin(α ± β) = sin α cos β ± cos α sin β

cos(−α) = cos α

sin(−α) =– sin α

to open Eq (5).

a n Rb
s(t) = cos [2πF c t + 2π t + Θn ]
4
a n Rb
= cos 2πF c t ⋅ cos (2π t + Θn ) –
4

Term 1

a n Rb
sin 2πF c t ⋅ sin (2π t + Θn )
4

Term 2

Now we process both terms one by one.

a n Rb
Term 1 = cos (2π t + Θn )
4
a n Rb a n Rb
= cos (2π t) cos Θn – sin (2π t) sin Θn
4 4
a n Rb Rb
= cos (2π t) cos Θn = dI ,n cos 2π t
4 4

because Θ0 = 0 implies sin Θn = 0, see Eq (7). Furthermore, a n  ϵ {−1, +1} and cos(±α) = cos α .

We have also defined

dI ,n = cos Θn

Similarly, Term 2 can be expanded as

a n Rb
Term 2 = sin (2π t + Θn )
4
a n Rb a n Rb
= sin (2π t) cos Θn + cos (2π t) sin Θn
4 4
Rb Rb
= cos Θn ⋅ a n sin 2π t = dQ,n sin 2π t
4 4

since sin Θn = 0 as before and sin(−α) =– sin(α). Here, we have defined dQ,n as

dQ,n = a n ⋅ cos Θn = a n ⋅ dI ,n
Finally, plugging both Term 1 and 2 into s(t) for the n-th bit period,

Rb Rb
s(t) = dI ,n cos 2π t ⋅ cos 2πF c t– dQ,n sin 2π t ⋅ sin 2πF c t
4 4

Using the identity cos α = sin(α + π/2), the above equation can be revised as

Rb π Rb
s(t) = dI ,n sin (2π t + ) ⋅ cos 2πF c t– dQ,n sin 2π t ⋅ sin 2πF c t
4 2 4
Rb
= dI ,n sin (2π (t + T b )) ⋅ cos 2πF c t–
4
Rb
dQ,n sin 2π t ⋅ sin 2πF c t —− Eq (8)
4
1
=dI ,n p(t + 2T b ) ⋅ cos 2πF c t– dQ,n p(t) ⋅ sin 2πF c t
2

where p(t) is a half-sinusoidal pulse shape of period 4T b .

Rb
p(t) = sin (2π t)
4

The above expression resembles an OQPSK waveform for bit n if the bit rate for dI ,n and dQ,n is Rb /2 or
bit period is 2T b , since the time offset in the sin term must be half the bit period for OQPSK. So we have to
check if dI ,n and dQ,n change values every other symbol.

From the definition, dI ,n = cos Θn and Eq (7) tells that Θn can only change values for odd n. Hence, dI ,n
is indeed an Rb /2 rate stream.

On the other hand, dQ,n = a n ⋅ dI ,n . Again, Eq (7) says that dI ,n can only change when a n changes but
that means that a n ⋅ dI ,n stays the same. Therefore, dQ,n can only change for even n and when a n
changes values. Consequently, dQ,n is also an Rb /2 rate stream.

Since dI ,n changes values for odd n while dQ,n does the same for even n, dQ,n is offset with respect to
Rb
dI ,n by T b seconds, the same amount as cos 2π t in Eq (8). Summing up everything so far, MSK can
4

indeed be represented as an OQPSK waveform. The data rate is the same as in CP-BFSK format since two
bits are being transmitted in two bit periods here as well.

This representation is illustrated in Figure 6.

 Figure 6: MSK as Offset QPSK

Compare Figure 6 with Figure 4. I did not choose separate blue and red colors in this figure so as not to
confuse dn s with a n s (that can raise a misunderstanding that dI ,n is odd a n s and dQ,n is even a n , which is
not correct).

Observe from Figure 6 that dI ,n is changing values every two T b s at odd multiples of T b , while dQ,n is
changing values every two T b s at even multiples of T b . Due to this offset behavior, at every T b , either I or
Q waveform is zero at t = Tb while the other reaches its maximum value. This is how phase remains
continuous during symbol transitions.

Figure 7 draws Θn as a supplement to above findings.

 Figure 7: Θn evolving with time

On a final note, observe from some equations (e.g., (2) and (5)) that the continuous phase has two parts, one
of which arises due to the delay of the n-th symbol. This information can be used to refine a phase estimate.

We are also clear now why MSK can act both like a linear and a non-linear modulation. In reality, MSK is a
non-linear modulation scheme (see Eq(2)) for a n . Pseudo-symbols dn themselves are non-linear functions of
information bits. So it is only from dn viewpoint that MSK can be seen as a linear modulation scheme.

A Generalization: From MSK to Continuous-Phase FSK (CP-FSK)

After understanding MSK, it can be expanded into a general modulation scheme known as Continuous-
Phase Frequency Shift Keying (CP-FSK). As we see in the next section, CP-FSK is a special case of
Continuous-Phase Modulation (CPM), which is a class of non-linear digital modulation schemes in which
the phase of the signal is constrained to be continuous from one symbol to the next. As with MSK, the most
attractive feature of such a signal is that it has a constant envelope as a result of the amplitude being
independent of the modulating information. Consequently, a CPM signal can be amplified without distortion by
a non-linear amplifier operating near the saturation point allowing low cost and more efficiency as compared
to a linear amplifier.

Usually CP-FSK (and hence CPM) is not a straightforward concept to master. However, starting from MSK, its
basics can easily be understood by building on the same expressions. For this purpose,

Remember that ΔF = R b /4.

Modulation symbols can carry more than 1 information bit. For example, 00, 01, 11 and 10 can be sent
through four symbols a n ϵ {−3, −1, +1 + 3}. In general, a n is a sequence from the alphabet
{±1, ±3, ⋯ , ±(M − 1)}. Thus, we call a n as symbols and replace T b with T M (a symbol time) from
here onwards. Symbol rate RM is then 1/T M .

Now let us start with rewriting Eq (2) and Eq (4) and then substituting the former into the latter.

n−1
π
θn = ∑ ai
2
i=0
 a n RM
s(t) = A cos [2πF c t + 2π (t − nT M ) + θn ]
4
n−1
1 t − nT M π
= A cos [2πF c t + 2π ⋅ an + ∑ ai]
2 2T M 2
i=0

n−1

= A cos [2πF c t + 2πha n ⋅ q(t − nT M ) + πh ∑ a i ]

i=0

where we have defined the following terms.

1
h =
2

⎧0 t ≤ 0
⎪
t
0 ≤ t ≤ TM
q(t) = ⎨ 2TM

⎩
⎪ 1
t ≥ TM
2

It turns out that h is a general concept known as

modulation index, which in case of MSK is equal to 1/2. In general, h is defined as

ΔF
h = 2
RM

which describes the peak frequency deviation in terms of a percentage of the symbol rate. With the definition
of h, we can write

n−1
1
s(t) = A cos [2πF c t + 2πh(a n q(t − nT M ) + ∑ a i )]
2
i=0

The above equation is true for the interval nT M ≤ t ≤ (n + 1)T M . Considering from its definition that
q(t) is 1/2 after a symbol interval, we can also write s(t) as

s(t) = A cos [2πF c t + 2πh ∑ a i q(t– iT M )] —− Eq (9)

i=0

t n

= A cos [2πF c t + 2πh ∫ ( ∑ a i g(u– iT M ))du]

−∞
i=0

where g(t) is defined as the derivative of q(t) as

q(t) = ∫ g(u)du

Notice that in the present case with q(t) defined as above, its derivative g(t) is a rectangular pulse shape.
Consequently, ∑ a i g(u– iT M ) is the standard baseband Pulse Amplitude Modulated (PAM) waveform with
i

rectangular pulse shape and this is how its discontinuities are transformed into a continuous-phase signal.
It is refreshing to conclude that the starting point for a continuous-phase modulated signal is a
standard pulse amplitude waveform, in which the discontinuities are smoothed out by the integral
operation.

More Generalization: From Continuous-Phase FSK (CP-FSK) to

Continuous-Phase Modulation (CPM)
Now referring to Eq (9), the expression for a CPM signal can be written as

s(t) = A cos [2πF c t + 2π ∑ hi ⋅ a i q(t– iT M )] nT ≤ t ≤ (n + 1)T

i=0

Here, hi is a sequence of modulation indices: a ratio of two relatively prime integers.

k
h =
p

This is the case of multi-h CPM. When all hi = h, the modulation index is the same for all symbols. This is
the category we saw in CP-FSK and MSK above.

Finally, q(t) is a waveform shape known as the phase response of the modulator which is normalized as

⎧ 0 t ≤ 0
⎪
t
q(t) = ⎨ 0 ≤ t ≤ LT
2T

⎩
⎪ 1
t ≥ LT
2

where L is the length of a pulse g(t) in symbols. Since angular frequency is the rate of change of phase, this
g(t) is the derivative of q(t) and known as the frequency response of the modulator.

dq(t)
g(t) =
dt

For any pulse shape g(t), L = 1 results in full response CPM, while the other case L > 1 is the partial
response CPM. Commonly used pulse shapes are rectangular (as used in the case of MSK), raised cosine
and Gaussian.

We can see that CPM is in fact a very large class of modulation schemes owing to different pulse shapes
g(t), modulation indices hi and the modulation alphabet size M . That is both a blessing and a curse:
blessing due to the remarkable variety of signals as its offsprings all yielding excellent spectral and power
efficiencies, and curse due to the high receiver complexity. By virtue of Moore’s law, this is becoming less of
an issue — thanks to powerful baseband processors in the modern age.

References
[1] Mengali and D’Andrea, Synchronization Techniques for Digital Receivers, 1997.

[2] John Proakis, Digital Communications, 4th ed, 2001.

 About the Author

Qasim Chaudhari is an independent consultant and trainer with expertise in wireless

communications, Software Defined Radio (SDR) and Digital Signal Processing (DSP).
He obtained a PhD in Electrical Engineering from Texas A&M University and worked
at various organizations including Qualcomm, Renzym, Center for Energy Efficient
Telecommunications and RMIT University. He has developed and implemented DSP
algorithms for real world demonstrations of many wireless systems such as a MIMO-
OFDM testbed, low-SNR receivers and phase of arrival based localization.

As the founder of https://wirelesspi.com, his focus is on explaining concepts through

beautiful figures, simple mathematics and intuitive reasoning. He is an author of two
books, Wireless Communications from the Ground Up - An SDR Perspective (along with
a video course with practical exercises to implement the concepts) and 5G Physical Layer
- An Easy Guide to Key Technologies. You can reach him at [email protected].