ELEC ENG 2104 Digital Signal Processing
Assignment 3 – Filter Design & Spectral Analysis
Semester 2, 2024
Instructions
Attempt all questions. Include Matlab plots in your submission. Upload a single pdf file on MyUni. No other formats will be accepted. The total mark for this assignment is 20. Part marks for each question are listed. Presentation will be taken into account in the marking.
This is an individual assignment. Discussing with others are encouraged but make sure you only submit work that is entirely your own. Including code templates and examples from MyUni is permitted but make sure these are appropriately referenced. Refer to the University’s Academic Integrity policy for details:
https://www.adelaide.edu.au/policies/230/
To complete this assignment, you need the file DSP2024_assigment3 .mat from MyUni. Put it in the same folder as your Matlab work files.
Question 1. A continuous time signal xa (t) has a spectrum (Fourier transform) shown in figure 1. This signal is to be sampled into a discrete time signal x[n]. Suggest an appropriate sampling rate for this situation. Briefly justify your answer.
Figure 1: Spectrum of continuous time signal to be sampled.
The DT signal is quantised using an ADC with a resolution of 12 bits. Estimate the signal to distortion ratio (SDR) of this quantisation scheme, stating any assumptions. [2 marks]
Question 2. A digital lowpass filter is needed to meet the following specifications:
• Passband edge ωp = 0.24π
• Stopband edge ωs = 0.3π
• Passband ripple δp = 0.2dB
• Stopband attenuation δs = 30dB
(a) Use the Matlab function fir1 to obtain a linear phase FIR windows-based design. [1 mark]
(b) Use the Matlab function firls to obtain a linear phase FIR least square error design. [1 mark]
(c) Use the Matlab function firpm to obtain a linear phase FIR Parks-McClellan (equiripple) design. [1 mark]
(d) Use the Matlab function ellip to obtain an elliptical IIR design. [1 mark]
(e) Briefly compare the orders of the designs and hence comment on their implementation costs. [1 mark]
For parts (a)–(d), show the impulse response, magnitude response and group delay plots to demonstrate your design meets the specifications. Clearly state any design parameters you have chosen.
Question 3. The complex-valued signal stored in the Matlab variable x3 consists of mulitple sinusoidal components and corrupted by white noise.
(a) Compute the DFT of the signal and plot the power spectral density (dB scale) as a function of frequency. From your spectrum, identify the frequency components which are present in the signal. The relevant Matlab command is fft. [2 marks]
(b) Now compute a windowed DFT of the signal using a Hamming window and plot the resultant energy density spectrum. Explain what you observe. Use the Matlab command window to generate the Hamming window. [2 marks]
Question 4. Sketch the signal flow graphs for the
(a) canonical implementation of the FIR filter with impulse response
h[n] ={1, −2, 5, −2, 1 }
[1 mark]
(b) first order parallel architecture of the IIR filter with transfer function
[1 mark]
(c) cascade implementation of the IIR filter with second order systems
[1 mark]
Question 5. (a) Calculate and write the magnitude (|X[k]|) and phase (7X[k]) of the 8-point DFT of
x[n] ={1, 2, 3, 1, 2, 3, 1, 2 } .
[1 mark]
(b) Consider the two DT signals
x1 [n] ={1, −2, −4, −1, 3, 2 } ,
x2 [n] ={−1, 5, 3, 2, −2, −3} .
Calculate the circular convolutions x1 [n] ⊛ x2 [n] of these two signals for (i) N = 6, (ii) N = 10 and (iii) N = 12. Compare these outputs with the linear convolution x1 [n] * x2 [n]. [2 marks]
(c) Use DFTs to calculate the output y[n] of the DT LSI system with impulse response h[n] and input x[n]
h[n] ={1, −1, 3, −1, 1 } ,
x[n] ={5, −8, 3, −1, 3, 6, −3, 8, −5, 0, 2, 5 } .
You can use convolution to check the answer, but it will not attract any marks. [1 mark]
Question 6. Consider the causal DT LSI system,
y[n] + 1.27y[n − 1] + 0.81y[n − 2] = x[n].
The system’s input x[n] is a Gaussian white noise process with variance 1.
(a) What is the autocorrelation of the output process RY Y [l] and power spectral density SYY (ω)? [1 mark]
(b) Simulate M = 500 realisations of white noise inputs, using N = 128 samples per realisation. Use the filter command to produce each output realisation and calculate its periodogram, i.e. squared magni- tude of its DFT. Average periodograms from all M realisations and compare the theoretical expectation from part (b). [1 mark]
Total: 20 marks
版权所有:编程辅导网 2021 All Rights Reserved 联系方式:QQ:99515681 微信:codinghelp 电子信箱:99515681@qq.com
免责声明:本站部分内容从网络整理而来,只供参考!如有版权问题可联系本站删除。