Professional Documents
Culture Documents
Part A
Sampling of Continuous-Time Signals
Part A: Sampling of Continuous-Time Signals Introduction Sampling of Continuous-Time Signals ContinuousEffect of Sampling in the Frequency Domain Recovery of the Analog Signal Implication of the Sampling Process
1. Introduction
Digital processing of a continuous-time signal involves the following basic steps: (1) Conversion of the continuous-time signal into a discrete-time signal, (2) Processing of the discrete-time signal, (3) Conversion of the processed discrete-time signal back into a continuous-time signal
1. Introduction
Conversion of a continuous-time signal into digital form is carried out by an analog-toanalog- todigital (A/D) converter The reverse operation of converting a digital signal into a continuous-time signal is performed by a digital-to-analog (D/A) digital- toconverter
1. Introduction
Since the A/D conversion takes a finite amount of time, a sample-and-hold (S/H) sample- andcircuit is used to ensure that the analog signal at the input of the A/D converter remains constant in amplitude until the conversion is complete to minimize the error in its representation To prevent aliasing, an analog anti-aliasing aliasing antifilter is employed before the S/H circuit.
1. Introduction
To smooth the output signal of the D/A converter, which has a staircase-like waveform, an analog reconstruction filter is used. The complete block-diagram is shown blew
Anti-aliasing filter S/H A/D Digital Processing D/A Reconstruction filter
1. Introduction
Since both the anti-aliasing filter and the reconstruction filter are analog lowpass filters, we review first the theory behind the design of such filters Also, the most widely used IIR digital filter design method is based on the conversion of an analog lowpass prototype
0.5
Amplitude
-0.5
-1
0.2
0.4
0.6
0.8
Time
Let ga(t) be a continuous-time signal that is sampled uniformly at t = nT, generating the sequence g(n) where g(n)= ga(nT) with T being the sampling period The reciprocal of T is called the sampling frequency FT , i.e., FT = 1/ T Now, the frequency-domain representation of ga(t) is given by its continuous-time Fourier transform (CTFT CTFT):
Ga ( j) = g a (t )e jt dt
The frequency-domain representation of g(n) is given by its discrete-time Fourier transform (DTFT DTFT): G (e j ) = g (n)e j n
n =
2T T
0 T 2T t
To establish the relation between Ga ( j) and G (e j ) , we treat the sampling operation mathematically as a multiplication of ga(t) by a periodic impulse train p(t):
ga(t) is a continuous-time signal consisting of a train of uniformly spaced impulses with the impulse at t = nT weighted by the sampled value ga(nT) of ga(t) at that instant
ga(t) gp(t) ga(t) 0
There are two different forms of G p ( j) : One form is given by the weighted sum of the CTFTs of (t nT ) : G p ( j) = g a (nT )e jnT
n =
To derive the second form, we note that p(t) can be expressed as a Fourier series: 1 + j 2 kt 1 + p ( t ) = e T = e jT kt T k = T k = where 2 T = T
The impulse train gp(t) therefore can be expressed as 1 + g p (t ) = e jT kt g a (t ) T k = From the frequency-shifting property of the CTFT, the CTFT of e jT kt g a (t ) is given by Ga ( j ( k T )) Hence, an alternative form of the CTFT of is given by 1 G p ( j) = Ga ( j ( k T )) T k =
Therefore, G p ( j) is a periodic function of consisting of a sum of shifted and scaled replicas of Ga ( j) , shifted by integer multiples of T and scaled by 1/T The term on the RHS of the previous equation for k = 0 is the baseband portion of G p ( j) , and each of the remaining terms are the frequency translated portions of G p ( j)
The frequency range T / 2 T / 2 is called the baseband or Nyquist band The above result is more commonly known as the sampling theorem: Let ga(t) be a bandlimited signal with Ga ( j) = 0 for > m ,then ga(t) is uniquely determined by its samples ga(nT), n ,if 2 T 2 m where T = T
1
m 0
G p ( j)
P ( j)
2T
3T
1/T
m
T m
T
2T
0 m
G p ( j )
2T
1/T
T m
T m m T T
2T
If T 2 m , ga(t) can be recovered exactly from gp(t) by passing it through an ideal lowpass filter H r ( j) with a gain T and a cutoff frequency c greater than m and less than T m as shown below
ga(t) gp(t)
The spectra of the filter and pertinent signals are shown below G ( j)
p
1/T
m
T m
T
2T
0 m T
H r ( j )
m < c < ( T m )
H r ( j)
g a (t )
c Ga ( j)
p(t)
1
m 0 m
On the other hand, if T < 2 m , due to the overlap of the shifted replicas of Ga ( j) , the spectrum Ga ( j) cannot be separated by filtering to recover because of the distortion caused by a part of the replicas Ga ( j) immediately outside the baseband folded back or aliased into the baseband. Several terms: T 2 m
Nyquist condition Folding frequency Nyquist frequency Nyquist rate Oversampling Undersampling Critical sampling
Note: A pure sinusoid may not be recoverable from its critically sampled version
Example 1 In high-quality analog music signal processing, a bandwidth of 20 kHz has been determined to preserve the fidelity Hence, in compact disc (CD music systems, CD) a sampling rate of 44.1 kHz, which is slightly higher than twice the signal bandwidth, is used
Example 2 Consider the three continuous time sinusoidal signals: g1 (t ) = cos(6 t ), g 2 (t ) = cos(14 t ), g3 (t ) = cos(26 t ) Their corresponding CTFTs are:
G1 ( j) = [ ( 6 ) + ( + 6 ) ] G2 ( j) = [ ( 14 ) + ( + 14 ) ] G3 ( j) = [ ( 26 ) + ( + 26 )]
6 0 G2 ( j)
14
14
G3 ( j)
These continuous-time signals sampled at a rate of T = 0.1 sec, i.e., with a sampling frequency T = 20 rad/sec The sampling process generates the continuous-time impulse trains, g1p(t), g2p (t), and g3p (t) Their corresponding CTFTs are given by
Glp ( j) = 10 Gl ( j ( k T )), 1 l 3
26
26
k =
10
H r ( j)
0.1 c 20 6 0 G2 p ( j) 0.1 c 20 14 6 0
G3 p ( j)
20
10 H r ( j)
c 14 20
We now derive the relation between the DTFT of g(n) and the CTFT of gp(t) To this end we compare
G (e j ) =
c = 10
with
10 H r ( j)
c
40
0.1
6 0
6
c
20
n =
g ( n )e
j n
20
40
The relation derived on the previous slide can be alternately expressed as 1 G (e jT ) = Ga ( j jk T ) T k = form G (e j ) = G p ( j) = / T or from G p ( j) = G (e j ) =T
It follows that G (e j ) is obtained from G p ( j) by applying the mapping = / T Now, the CTFT G p ( j) is a periodic function of with a period T = 2 / T Because of the mapping, the DTFT G (e j ) is a periodic function of with a period 2
We now derive the expression for the output ga ( fsdft ) of the ideal lowpass reconstruction filter H r ( j) as a function of the samples g(n) The impulse response hr(t) of the lowpass reconstruction filter is obtained by taking the inverse DTFT of : T , c H r ( j ) = 0, > c
Thus, the impulse response is given by 1 T c jt j t hr (t ) = H r ( j)e d = 2 c e d 2 sin( c t ) , t = ct / 2 The input to the lowpass filter is the impulse train gp(t) g p (t ) = g ( n) (t nT )
n =
g(n)h (t nT )
r
Substituting hr (t ) = sin( ct ) / c t / 2 in the above and assuming c = T / 2 = / T for simplicity, we get sin[ (t nT ) / T ] g a (t ) = g ( n) (t nT ) / T n =
It can be shown that when c = T / 2 in sin c t hr (t ) = T t / 2 hr(0)=1 and hr(nT)=0 for n0 As a result, from sin[ (t nT ) / T ] g a (t ) = g ( n) (t nT ) / T n = we observe g a (rT ) = g (r ) = g a (rT ) < r < for all integer values of r in the range
The relation g a (rT ) = g (r ) = g a (rT ) holds whether or not the condition of the sampling theorem is satisfied However, g a (t ) = g a (t ) for all values of t only if the sampling frequency T satisfies the condition of the sampling theorem
Consider again the three continuous-time signals: g1 (t ) = cos(6 t ) , g 2 (t ) = cos(14 t ) , and g3 (t ) = cos(26 t ) The plot of the CTFT G1 p ( j) of the sampled version of g1(t) is shown below
G1 p ( j)
0.1
10 H r ( j)
20
c 6
20
From the plot, it is apparent that we can recover any of its frequency-translated versions cos ( 20k 6 ) t outside the baseband by passing through an ideal analog bandpass filter g1p(t) with a passband centered at = ( 20k 6 ) For example, to recover the signal cos(34 t ) cos(34pt), it will be necessary to employ a bandpass filter with a frequency response 0.1, (34 ) (34 + ) H r ( j ) = A small number otherwise 0,
Likewise, we can recover the aliased baseband component cos(6 t ) from the cos(6pt) sampled version of either g2p(t) or g3p(t) by passing it through an ideal lowpass filter with a frequency response:
0.1, 0 (6 + ) H r ( j ) = otherwise 0,
There is no aliasing distortion unless the original continuous-time signal also contains the component cos(6 t ) Similarly, from either g2p(t) or g3p(t) we can recover any one of the frequency-translated versions, including the parent continuous-time signal cos(14t) or cos(26t) as the case may be, by employing suitable filters
This leads to G p ( j) =
0
G p ( j)