complex Gaussian noise signals with zero mean and different variance?

Hi;

 I need to generate in matlab  complex Gaussian noise signals with zero mean and different typical values of variance  .

Answer 1

[Andrew Newell]

Assuming your signals are not correlated, you could do it with

cgnoise = std1*randn(n,1) + i*std2*randn(n,1);

where std1 and std2 are the standard deviations.

(Edited to correct errors)

[zayed]

what a bout implementing the following ,in aloop

cgnoise=sqrt(var)*randn(n,1)+j*sqrt(var)*randn(n,1)

[Andrew Newell]

If you want to generate one value at a time, you could use

for ii=1:n

cgnoise = std1*randn + i*std2*rand;

end

[Mohammad Alsaleh]

why you use different std for the real and imaginary part?

Are they different in reality for thermal noise ?

[Sarah Crimi]

I am not sure about the std1 vs std2 but, it should be cgnoise(ii) in the loop so that it obtains a value for each ii.

for ii=1:n

cgnoise(ii) = std1*randn + i*std2*rand;

end

[Srikanth V]

Does not make sense to have diff std_dev for real and imag..if you want snr controlled noise.,

var = (N different var);

for lp_var=1 : size(var,2)

noise_lvl(lp_var) = sqrt(lp_var)*complex(rand(1,sig_len),rand(1,sig_len));

end