DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING NATIONAL INSTITUTE OF TECHNOLOGY TRICHY DIGITAL SIGNAL PROCESSING LAB (EC315) 2009‐2010 BATCH CYCLE 1 USING MATLAB EXPERIMENT 1 STUDY OF CORRELATION OPERATION 1) Generate the signal of the form ]
2) Note that the individual signals are appended. 3) Generate Y
.
4) Correlate the signal Y[n] with X[n] to obtain Z[n]. (what do you infer?) 5) Repeat the experiment with . • Y • Y . • Y . EXPERIMENT 2 STUDY OF CONVOLUTION OPERATION 1) Generate the signal of the form S[n]= ∑ sin 2 Π 2) Convolve the signal with the filter whose impulse response is given below to obtain y1 [n]. h[n] =[‐0.0011 0.0048 0.0006 ‐0.0316 0.0275 0.0975 ‐0.1298 ‐0.2263 0.3153 0.7511 0.4946 0.1115]
3) Replace the alternative samples of y1[n] with zeros to obtain y2[n]
4) Convolve the signal with the filter whose impulse response is given below to obtain y3[n]. h[n] =[ ‐0.1115 0.4946 ‐0.7511 0.3153 0.2263 ‐0.1298 ‐0.0975 0.0275 0.0316 0.0006 ‐0.0048 ‐0.0011]
5) Replace the alternative samples of y3[n] with zeros to obtain y4[n]
6) Convolve the signal y2[n] with the filter whose impulse response is given below to obtain out1[n] h[n]=[ 0.1115 0.4946 0.7511 0.3153 ‐0.2263 ‐0.1298 0.0975 0.0275 ‐0.0316 0.0006 0.0048 ‐0.0011]
7) Convolve the signal y4[n] with the filter whose impulse response is given below to obtain out2[n] h[n]= [‐0.0011 ‐0.0048 0.0006 0.0316 0.0275 ‐0.0975 ‐0.1298 0.2263 0.3153 ‐0.7511 0.4946 ‐0.1115]
8) Use subplot to plot the following signals • S[n] • Y1[n] • Y3[n] • out1[n]+out2[n] What do you infer? Note: 1. Use Overlap and add method to obtain the output 2. Realize the Linear convolution using Circular convolution 3. Realize the circular convolution using DFT technique. EXPERIMENT 3 STUDY OF FILTERING USING FIR FILTER 1) Generate the signal of the form sin 2 1 sin 2 2 sin 2 3 Such that and 3 1 2 as the noise .Design and Realize the Linear 2) Assume the signal sin 2 3 phase FIR filter using hamming window to filter the noise to obtain the output signal Y[n]. 3) Use subplot to plot the following signals sin 2 1 sin 2 21 • • X[n] • Y[n] 4) Draw the Magnitude and phase response of the designed filter 5) Draw the magnitude and Phase response of the signal before and after filtering. 6) Repeat the experiment with the following windows for designing FIR filter. • Rectangular window • Barlett window • Hanning window
• •
Hamming window Blackman window 7) Plot the time domain sequence (n=0,1,2…M‐1) and the corresponding frequency response plot of the following windows for M=31 8) Measure the following frequency domain characteristics of the above mentioned windows. • Transition width of the main lobe • Peak side lobe EXPERIMENT 4 Part‐1 STUDY OF FILTERING USING IIR FILTER Generate the signal of the form sin 2
1)
2)
3) 4) 5)
1
sin 2 1
2
sin 2
3
Such that and 3 2 as the noise .Design and Realize the IIR filter 3 Assume the signal sin 2 using Bilinear transformation to filter the noise to obtain the output signal Y[n]. Use subplot to plot the following signals • sin 2 1 sin 2 21 • X[n] • Y[n] Plot the magnitude and phase response of the designed filter Draw the magnitude and Phase response of the signal before and after filtering. Repeat the experiment with Impulse in‐variant transformation for designing IIR filter.
Part‐2 STUDY OF INTERPRETATION OF POLE‐ZERO PLOT Draw the magnitude and Phase response of the digital filter whose pole‐zero plot is as shown below.
EXPERIMENT 5 STUDY OF FINITE WORD LENGTH EFFECTS 1) (a) Load the signal signal.dat .Note that the length of the signal is 8000 and is stored in the variable ‘x’ with sampling frequency 16000Hz. (b) Divide the signal ‘x’ into group of samples with size 512 (c) Compute 512 point fft using inbuilt function fft.m for every group and store it in the variable ‘y’ (d)Divide the signal ‘y’ into groups of samples with size 512 (e)Compute 512 point ifft using radix‐2 DIT FFT algorithm (your function) for every group with fixed point operation with the following specifications and store in the variable ‘z’ ¾ Fractional length of the bits=4 bits ¾ Total length of the bits=16+3=7 bits (f) Compute the SQNR (Signal to Quantization Noise) using the formula 20*log10 (sum (x .^2)/sum ((x‐z) .^2)) (g)Repeat the above experiment with the following fractional bit lengths ranging from 5 to 18. (h) plot the graph between fractional bit lengths and the corresponding SQNR (i)Repeat the above experiment by collecting the samples with size 1024. [Use the inbuilt function fi.m for fixed point operation. But you must know how to represent the number using fixed point representation and how to compute (add / mul) using fixed point representation]
General Guidelines 1. Write Aim, Procedure with example ,Program, Report in your observation 2. Writ Aim, Description, Procedure with example, Program, Graphs, Report in your Record. 3. All the Graphs should have sampling time/Sampling frequency 4. X‐axis in all the graphs should be in seconds in time domain 5. X‐axis in all the graphs should be in Hz in frequency domain 6. Graphs should be plotted with White background color with black lines 7. Zoom the figure wherever required before taking the printouts 8. Don’t use inbuilt function related to Signal processing. Use the same for cross checking 9. Name the file name with arbitrary name followed by two digit roll number of the first person of the batch. 10. Type help followed by the function name in the command window to read about that function. 11. Avoid for‐loops as for as possible 12. Reports should have all the interpretations you have made on the particular experiment 13. All the Equations, texts and picture must be clear and neat in both the record and observation. ALL THE BEST