How Do I Process AR and Speech Signals in MATLAB?

[ahmedxx]

matlab help!

I have a task :

Two different signals will be used as input signal u . The first one is an AR model generated
signal. The second one is a speech signal recorded with 16bits/sample at 8kHz sampling rate. The name of the speech file is speech.mat

**Generating an order 40 AR signal:

i. Generate randomly the poles of the model and find the coefficients of polynomial from the
roots (Matlab function poly).>> (I have done this)

ii. Generate 25000 samples according to this model (unitary variance driving noise).

iii. Find the “true” autocorrelation matrix starting from the coefficients of the model. Let us
write the Yule-Walker equations having the values of autocorrelation function as
unknowns.


Can anybody help me and which command I can use?

 

[Valkarie]

To generate 25000 samples according to this model, you can use the Matlab command "filter". The syntax is:y = filter(B,A,u)where B is the numerator polynomial of your filter and A is the denominator polynomial. u is a vector of noise with unitary variance. To find the “true” autocorrelation matrix, you can use the Matlab command "xcorr". The syntax is: [R, lags] = xcorr(x, maxlag) where x is the vector containing the AR model generated signal and maxlag is the maximum lag for which you want to compute the autocorrelation. Finally, to write the Yule-Walker equations, you can use the Matlab command "ar2rc". The syntax is: [rc, lag] = ar2rc(arcoef, maxlag) where arcoef is the vector containing the coefficients of the model and maxlag is the maximum lag for which you want to compute the autocorrelation.

 

[Mene]

Sure, I would be happy to help you with your MATLAB signal processing task. To generate an AR model signal, you can use the "ar" function in MATLAB. This function takes in the coefficients of the polynomial and the noise variance as inputs and generates the desired signal. For example, if your coefficients are stored in a vector called "a" and the noise variance is stored in a variable called "var", the command would be:

u = ar(a, var, 25000);

To find the true autocorrelation matrix, you can use the "xcorr" function which calculates the cross-correlation between two signals. In this case, you can use it to calculate the autocorrelation of the signal "u" that you generated and compare it with the autocorrelation function that you obtained from the Yule-Walker equations.

I hope this helps. Best of luck with your task!