finding phase using autocorrelation

[ashwini1]

Hi,

I have a 30Hz wave, where i have introduced a phase shift of 45 degrees. Somewhere i read that the phase info can be computed based on auto correlation function. I tried and i did not get required output. Here is my code, can anyone tell me that what is wrong in this.

Fs=100e3;
t=[0:Fs]/Fs;
x=sin(2*pi*30*t+(45*pi/180));
c=xcorr(x);
t1=[-100e3:100e3];
plot(t1,c);

[val,index]=max(c);
lag=t1(index);
ph=lag*180/pi;

i suppose to get ph=45; instead i am getting 0.

Any help plz ..
Thanks,
Ashwini

 

[MHanif]

Hi

The following code will give you the phase offset.

Fs=100e3;
t=[0:Fs]/Fs;
x=sin(2*pi*30*t+(15*pi/180));
y=sin(2*pi*30*t);
c=xcorr(y,x);
t1=[-100e3:100e3]/Fs*2*pi*30;
plot(t1,c);

[val,index]=max(c);
lag=t1(index);
ph=lag*180/pi;

 

[ashwini1]

Thanks for your reply,

can you plz explain me this:
t1=[-100e3:100e3]/Fs*2*pi*30;

why are you diving t1 by the factor:Fs*2*pi*30.

Is it to change the units from sec to radians.

Thanks
Ashwini

 

[MHanif]

Hi Ashwini

why are you diving t1 by the factor:Fs*2*pi*30.

Firstly I am not dividing by Fs*2*pi*30.
I am dividing by Fs and multiplying by 2*pi*30.

Division by Fs makes 't1' time axis i-e if you observe that division by Fs makes the variable 't1' having dimensions of time.

Then I multiplied it with 2*pi*30 to convert this time lag into phase lag!

This time give me Thanks by clicking Helped me

 

[ashwini1]

Thanks a lot