adding 5% white gaussian noise

Hi everyone;

I need to add 5% noise to my signal (amplitude of noise = 0.05* signal amplitude) using "randn" as:

signal_noisy=signal+0.5*randn(signal) % (is it correct?)

is the genereated noise, whight gussion one?!

if I want to use "awgn" is it correct to say in this case SNR=13 ???

I realy appriciate your help here

Answer 1

[Image Analyst]

Try this:

% Make sample data.
x = linspace(0, 3*pi, 640);
period = 2*pi;
y = cos(2 * pi * x / period); % Noise-free signal.
% Noise is 5% of the signal value at each location, 
% which means sigma is not a constant at each location.
amplitudes = y * 0.05;
noise = amplitudes .* randn(1, length(y));
% Add the noise to the noise-free signal.
yNoisy = y + noise;
% Plot everything.
plot(x, yNoisy, 'r.', 'MarkerSize', 8);
hold on;
plot(x, y, 'b-', 'LineWidth', 2);
grid on;

[atabak mostafavi]

great thank you; and would this method generate gussion ( flat spactrum ) noise since I am trying to model the measurement noise and it kinda important to be flat spactrum.

[Image Analyst]

Yes, it's flat. The noise depends only on the signal and there is no spatial or time dependency on it, other than the signal itself of course.

[atabak mostafavi]

thank you