Department of Electronic Systems

Examination paper for

TTT4120 Digital Signal Processing

Academic contact during examination: Stefan Werner

Phone: +358401424948

Examination date: Tuesday, August 7, 2018

Examination time (from-to): 09:00-13:00
Permitted examination support material:
D - Basic calculator allowed
No printed or handwritten materials allowed
Other information:
 Exam consists of four (4) problems
 A few basic formulas are provided in the Appendix

Language: English.
Number of pages (front page excluded): 6
Number of pages enclosed: 2

Informasjon om trykking av eksamensoppgave

Checked by:
Originalen er:

1-sidig □ 2-sidig □ ____________________________

sort/hvit □ farger □ Date Signature

skal ha flervalgskjema □

___________________________________________________________________________________________
Merk! Studentane finn sensur i Studentweb. Har du spørsmål om sensuren må du kontakte instituttet ditt.
Eksamenskontoret vil ikkje kunne svare på slike spørsmål.
Problem 1 (9+5+6+3=23): Basics of filter theory and design

1a) Consider a discrete-time system with input-output relation 𝑦[𝑛] = ℋ{𝑥[𝑛]}, where operator
ℋ{∙} describes the system. Below you find four different discrete-time systems

i. 𝑦[𝑛] = (𝑛 + 1)𝑥[𝑛]
ii. 𝑦[𝑛] = 𝑥[𝑛] + 0.2𝑦[𝑛 − 1]
iii. 𝑦[𝑛] = √𝑥[𝑛 + 1], with 𝑥[𝑛] ≥ 0 ∀𝑛
iv. 𝑦[𝑛] = 0.3𝑥[𝑛] + 2

 Which properties need to be fulfilled in order to fully describe a system using its
impulse response ℎ[𝑛]? List which systems above, i.-iv., have these properties? (3p)
 Define the system property causality in terms of ℎ[𝑛]. List which systems above, i.-iv.,
are causal. (3p)
 Define the system property stability in terms of ℎ[𝑛]. List which systems above, i.-iv.,
are stable. (3p)

1b) Assume knowledge of the impulse response ℎ[𝑛], ∀𝑛 ∈ (−∞, ∞).

 Define the z-transform 𝐻(𝑧) (1p)

 Sketch the ROC in the z-plane for a causal and for an anti-causal system (2p)
 What must belong to the ROC if the system is to be stable? (2p)

1c) Consider a filter with system function

1 + 𝑏𝑧 −1
𝐻(𝑧) = , 𝑎, 𝑏 ≥ 0
1 − 𝑎𝑧 −1
Give the possible values of the coefficients {𝑎, 𝑏} and the corresponding ROCs when the
system 1) is stable, 2) is causal, and 3) has minimum phase.

1d) Derive the expression for the impulse response ℎ[𝑛] = 𝒵 −1 {𝐻(𝑧)} for 𝐻(𝑧) given in 1c).

1 (4)
Problem 2 (6+6+6+2=20): Filter structures and implementations

Consider the system

𝐻(𝑧) = 𝐻1 (𝑧)𝐻2 (𝑧)𝐻3 (𝑧)

1 1 −1 1 −1
where 𝐻1 (𝑧) = 1 − 2 𝑧 −1 , 𝐻2 (𝑧) = (1 − 4 𝑧 −1 ) , and 𝐻3 (𝑧) = (1 + 4 𝑧 −1 ) .

The system is to be implemented using fixed-point representation with 𝐵 + 1 bits (one bit is used
for the sign) and dynamic range [−1,1). Rounding is performed after each multiplication and the
rounding error 𝑒[𝑛] can be modeled as white noise with variance 𝜎𝑒2 . Consequently, each
multiplier in the fixed-point implementation is modeled as
𝑄(𝑎𝑦[𝑛 − 𝑘]) = 𝑎𝑦[𝑛 − 𝑘] + 𝑒[𝑛]

which is equivalent to adding noise sources after multipliers in the infinite-precision realization.
Rounding noise sources combine into an equivalent noise signal 𝑧[𝑛] at the filter output with
variance 𝜎𝑧2 (see hint below).
2a) Draw the cascade structure, of first-order sections, 𝐻(𝑧) = 𝐻1 (𝑧)𝐻2 (𝑧)𝐻3 (𝑧), with noise
sources due to rounding included. Determine the variance of the round-off noise at the filter
output.

2a) Draw the cascade-structure, 𝐻(𝑧) = 𝐻1 (𝑧)𝐻2′ (𝑧), with noise sources due to rounding
included, where filter 𝐻2′ (𝑧) = 𝐻2 (𝑧)𝐻3 (𝑧) is implemented as a second-order section with
real coefficients. Determine the variance of the round-off noise at the filter output.

2c) Draw the equivalent parallel-structure, 𝐻(𝑧) = 𝐻4 (𝑧) + 𝐻5 (𝑧), with noise sources due to
rounding included. Determine 𝐻4 (𝑧) and 𝐻5 (𝑧), and the variance of the round-off noise at
the filter output.

2d) How many possible cascade realizations, of first-order sections exist?

2
[Hint:] Assuming noise source 𝑒𝑖 [𝑛] with variance 𝜎𝑒𝑖 acts as input to (sub-)filter ℎ𝑖 [𝑛] that
terminates at the output, the variance of the noise signal 𝑧𝑖 [𝑛], due to 𝑒𝑖 [𝑛], is given by
1
2
𝜎𝑧𝑖 2
= 𝜎𝑒𝑖 2
𝑟ℎ𝑖ℎ𝑖 [0] = 𝜎𝑒𝑖 ∑ ℎ𝑖2 [𝑘] = 𝜎𝑒𝑖
2
∫ |𝐻𝑖 (𝑓)|2 𝑑𝑓
𝑘 0

𝑒𝑖 [𝑛] 𝑧𝑖 [𝑛]
𝐻𝑖 (𝑧)

2
𝜎𝑒𝑖 = 𝐸{𝑒𝑖2 [𝑛]} 2
𝜎𝑧𝑖 = 𝐸{𝑧𝑖2 [𝑛]}

2 (4)
Problem 3 (4+8+8=20): Parametric modeling and Wiener filtering

𝑊[𝑛] 𝑆[𝑛]
𝐻(𝑧)

𝛾𝑊𝑊 [𝑙] 𝛾𝑆𝑆 [𝑙]

Γ𝑊𝑊 (𝑓) Γ𝑆𝑆 (𝑓)

Fig. 1: Filtering of stochastic processes

A wide-sense stationary (WSS) stochastic process 𝑆[𝑛] is generated by filtering a white noise
process 𝑊[𝑛], with autocorrelation sequence 𝛾𝑊𝑊 [𝑙] = 0.36𝛿[𝑙], through a causal and stable
filter with the following system function

1
𝐻(𝑧) =
1 − 0.8𝑧 −1

3a) Design the whitening filter 𝐻𝐼 (𝑧) so that the cascaded structure 𝐻(𝑧)𝐻𝐼 (𝑧) produces a white
output sequence, when 𝑊[𝑛] is the input sequence. Provide the resulting ROC for 𝐻𝐼 (𝑧).

3b) Compute the output spectrum Γ𝑆𝑆 (𝑓) of signal 𝑆[𝑛] and show/verify that the corresponding
autocorrelation sequence is given by 𝛾𝑆𝑆 [𝑙] = 0.8|𝑙| .

3c) The signal 𝑆[𝑛], with known 𝛾𝑆𝑆 [𝑙], is then corrupted by additive white noise resulting in
another (noisy) sequence 𝑋[𝑛] = 𝑆[𝑛] + 𝑉[𝑛], with 𝑉[𝑛] being uncorrelated white noise
with power 𝜎𝑉2 = 1 which is independent of 𝑆[𝑛]. We wish to filter the noisy signal 𝑋[𝑛]
using an FIR Wiener filter with length 𝑀 = 2. Find the coefficients of the FIR filter, i.e., the
filter that minimizes 𝜎𝐸2 = 𝐸{(𝑆[𝑛] − ∑𝑀−1 2
𝑘=0 ℎ[𝑘]𝑋[𝑛 − 𝑘]) }.

Hint: Remember that the inverse of a 2-by-2 matrix is given by the following formula

𝑎 𝑏 −1 1 𝑑 −𝑏
( ) = ( )
𝑐 𝑑 𝑎𝑑 − 𝑏𝑐 −𝑐 𝑎

3 (4)
Problem 4 (4+16=20): Sampling, filtering, and rate-conversion

𝑡𝑛 = 𝑛𝑇𝑥
𝑥𝑎 (𝑡) 𝑥[𝑛] 𝑥̃[𝑛] 𝑣[𝑘] 𝑤[𝑘] 𝑦[𝑚]
𝑔[𝑛] ↑𝐼 ℎ[𝑘] ↓𝐷

𝐼
Rate: 𝐹𝑥 Rate: 𝐹𝑦 = 𝐹𝑥
𝐷

𝐼
Fig. 2: Rate-conversion by a fractional factor 𝐷

The continuous-time signal 𝑥𝑎 (𝑡) = 2 cos 400𝜋𝑡 + 4 cos 800𝜋𝑡 is sampled at every 𝑇𝑥 second
to generate the discrete-time sequence 𝑥[𝑛] = 𝑥𝑎 (𝑡)|𝑡 =𝑛𝑇𝑥 .

4a) What is the minimum sampling rate, or the Nyquist rate, that ensures that 𝑥𝑎 (𝑡) can be
reconstructed from samples 𝑥[𝑛]? Plot the spectrum of the discrete-time sequence 𝑥[𝑛]
obtained by sampling 𝑥𝑎 (𝑡) at Nyquist rate, i.e., 𝑋(𝑓) = ∑∞
𝑛=−∞ 𝑥[𝑛]𝑒
−𝑗2𝜋𝑓𝑛
with 𝑓 = 𝐹𝑇𝑥
being the normalized frequency.

4b) Signal 𝑥𝑎 (𝑡) is now sampled at every 𝑇𝑥 = 0.625 ms. You want to remove the high-
frequency component and, thereafter, change the sampling rate in digital domain to 𝐹𝑦 =
1200 Hz. Figure 2 shows the schematic that will perform this task. Filter
ℎ[𝑘] is a lowpass filter with frequency response
1
1, |𝑓𝑣 | ≤ 2 max(𝐼,𝐷)
𝐻(𝑓𝑣 ) = {
0, otherwise

 Assume that filter 𝑔[𝑛] is an ideal lowpass filter. Provide the range of frequencies
that can be used as valid cutoff frequency so that 𝑥̃[𝑛] only contains the low-
frequency sinusoidal sequence. (2p)
 Is the sequence 𝑥̃[𝑛] periodic or aperiodic? Motivate. (2p)
 Provide the smallest integer values of 𝐼 and 𝐷 needed for this rate conversion? (2p)
 Sketch the spectra of signals 𝑥̃[𝑛], 𝑣[𝑘], 𝑤[𝑘] and 𝑦[𝑚]. (8p)
 Comment on whether any information is lost in the rate conversion. Motivate. (2p)

4 (4)
Appendix: TTT4120 Table of formulas, 2018

A. Sequences:

1−𝛼𝑁
∑𝑁−1 𝑛
𝑛=0 𝛼 = 1−𝛼

1 1
|𝛼| < 1 ⇒ ∑∞ 𝑛
𝑛=0 𝛼 = 1−𝛼 and − ∑−∞ 𝑛
𝑛=−1 𝛼 = 1−𝛼

1−𝛼𝑁 𝑁𝛼𝑁
∑𝑁−1 𝑛
𝑛=0 (𝑛 + 1)𝛼 = (1−𝛼)2 − 1−𝛼 ; 𝛼≠1

1
|𝛼| < 1 ⇒ ∑∞ 𝑛
𝑛=0(𝑛 + 1)𝛼 = (1−𝛼)2

B. Linear convolution:

𝑦[𝑛] = ℎ[𝑛] ∗ 𝑥[𝑛] = ∑∞ ∞

𝑘=−∞ ℎ[𝑘]𝑥[𝑛 − 𝑘] = ∑𝑘=−∞ 𝑥[𝑘]ℎ[𝑛 − 𝑘]

𝑌(𝑧) = 𝐻(𝑧)𝑋(𝑧)

𝑌(𝑓) = 𝐻(𝑓)𝑋(𝑓)

𝑌(𝑘) = 𝐻(𝑘)𝑋(𝑘), 𝑘 = 0,1, … , 𝑁 − 1 where 𝑌(𝑘) = 𝑌(𝑓𝑘 ) with 𝑓𝑘 = 𝑘/𝑁

C. Transforms:

Z-transform: 𝐻(𝑧) = ∑∞
𝑛=−∞ ℎ[𝑛]𝑧
−𝑛

DTFT: 𝐻(𝑓) = ∑∞
𝑛=−∞ ℎ[𝑛]𝑒
−𝑗2𝜋𝑓𝑛

DFT: 𝐻(𝑘) = ∑𝑁−1

𝑛=0 ℎ[𝑛]𝑒
−𝑗2𝜋𝑛𝑘/𝑁
, 𝑘 = 0,1, … , 𝑁 − 1
1
IDFT: ℎ[𝑛] = 𝑁 ∑𝑁−1
𝑘=0 𝐻(𝑘)𝑒
𝑗2𝜋𝑛𝑘/𝑁
, 𝑛 = 0,1, … , 𝑁 − 1

D. Sampling theorem:

Given an analog signal 𝑥𝑎 (𝑡) sampled at 𝐹𝑠 = 1/𝑇. The DTFT of the resulting discrete-time
sequence 𝑥[𝑛] = 𝑥𝑎 (𝑡)|𝑡=𝑛𝑇 is given by

𝑋(𝑓) = 𝑋(𝐹/𝐹𝑠 ) = 𝐹𝑠 ∑∞
𝑘=−∞ 𝑋([𝑓 − 𝑘]𝐹𝑠 )

E. Autocorrelation, energy spectrum and Parseval:

Given a sequence ℎ[𝑛] with finite energy 𝐸ℎ

Autocorrelation: 𝑟ℎℎ [𝑙] = ∑∞

𝑛=−∞ ℎ[𝑛]ℎ[𝑛 + 𝑙] 𝑙∈ℤ

Energy spectrum: 𝑆ℎℎ (𝑧) = 𝐻(𝑧)𝐻(𝑧 −1 ) ⇒ 𝑆ℎℎ (𝑓) = |𝐻(𝑓)|2

1
Parseval’s theorem: 𝐸ℎ = 𝑟ℎℎ [0] = ∑∞ 2 2
𝑛=−∞ ℎ [𝑛] = ∫0 |𝐻(𝑓)| 𝑑𝑓

Appendix: TTT4120 Table of formulas A1 (A2)

F. Multirate:

Decimation (downsampling) where 𝑇𝑦 = 𝐷𝑇𝑥

𝑣(𝑚𝑇𝑦 ) = ∑∞
𝑘=−∞ ℎ[(𝑚𝐷 − 𝑘)𝑇𝑥 ]𝑥(𝑘𝑇𝑥 ) 𝑚∈ℤ

Interpolation (upsampling) where 𝑇𝑦 = 𝑇𝑥 /𝐼

𝑦(𝑙𝑇𝑦 ) = ∑∞
𝑛=−∞ ℎ[(𝑙 − 𝑛𝐼)𝑇𝑦 ]𝑥(𝑛𝑇𝑥 ) 𝑙∈ℤ
𝐷
Rate conversion where 𝑇𝑦 = 𝐷𝑇𝑣 = 𝐼 𝑇𝑥

𝑦(𝑙𝑇𝑦 ) = ∑∞
𝑚=−∞ ℎ[(𝑙𝐷 − 𝑚𝐼)𝑇𝑣 ]𝑥(𝑚𝑇𝑥 ) 𝑙∈ℤ

G. Autocorrelation, power density spectrum and Wiener-Khintchin:

Given a wide-sense stationary and ergodic sequence 𝑋[𝑛] with infinite energy

Autocorrelation: 𝛾𝑋𝑋 [𝑙] = 𝐸{𝑋[𝑛]𝑋[𝑛 + 𝑙]} 𝑙∈ℤ

Power spectrum: 𝛤𝑋𝑋 (𝑧) = 𝒵{𝛾𝑋𝑋 [𝑙]}

Wiener-Khintchin: 𝛤𝑋𝑋 (𝑓) = DTFT{𝛾𝑋𝑋 [𝑙]} = ∑∞

𝑙=−∞ 𝛾𝑋𝑋 [𝑙]𝑒
−𝑗2𝜋𝑓𝑙

H. Yule-Walker and Normal equations where 𝒂𝟎 = 𝟏:

Autocorrelation: ∑𝑃𝑘=0 𝑎𝑘 𝛾𝑋𝑋 [𝑛 − 𝑘] = 𝜎𝑓2 𝛿[𝑛] 𝑛 = 0, … , 𝑝

Normal equations: ∑𝑃𝑘=1 𝑎𝑘 𝛾𝑋𝑋 [𝑛 − 𝑘] = −𝛾𝑋𝑋 [𝑛] 𝑛 = 1, … , 𝑝

H. FIR Wiener filter design:

MSE filter: ∑𝑀−1

𝑘=0 ℎ[𝑘]𝛾𝑋𝑋 [𝑛 − 𝑘] = 𝛾𝑆𝑋 [𝑛] 𝑛 = 0, … , 𝑀 − 1

J. Some common z-transform pairs:

Appendix: TTT4120 Table of formulas A2 (A2)