$30.00
Order NowYou are given x.wav and v2.wav (both 74000 samples long at 32kHz). x[n] is the sum of a speech
signal s[n] (yes, really, there’s a speech signal in there) plus additive noise v1[n]. v2[n] is a signal which
is highly correlated with v1[n] (the noise in x[n]), but does not contain any of the speech. Your job is to
recover the clean speech. You may use any approaches covered in this class and any built-in Matlab
functions.
Design filters of order 4 through 20 in steps of 4. For each filter you design, provide some numerical
results about the quality of the filter. Also, comment on the perceptual quality of the output signal (i.e.
how your estimate of s[n] sounds when you listen to it) over the various filters. Finally, give a guess
about the system architecture and filter order(s) which produced the given signals.
Turn in
• a 1-page written summary of your approach (i.e. 1 page of text, possibly including a couple of
equations if appropriate);
• the coefficients and/or formulas for all the filters you design, along with numerical/graphical
results for each of them;
• the other results mentioned above;
• all your code.
Your report will be graded on correctness of the approach(es) you used, and completeness of the
required elements of the assignment. You will not be penalized strictly on the basis of your results if the
methods you used were appropriate and you addressed all the requirements. You are encouraged to
discuss your approaches with other students. However, do not share code: since this assignment is
unique and there are many correct approaches, it will be very easy to spot copied code.
•
WhatsApp us