Hi, I think what you are looking for here is rms power. Therefore you will have: normalized_noise=noise/std(noise,1); noise_power = reqdpower*normalized_noise; Generally you dont find the power of the entire segment all at once, but since this is gaussian noise, it should not really matter. Mandar --- nammi sairamesh <> wrote: > > Hi, > Multiply by sqrt(N) to the ranom smaples u > generated. > I mean sqrt(N)*randn will generate a random noise with > desied variance of N. > HTH, > -SaiRamesh. > --- sharma rohit <> wrote: > > > Hi! Folks > > I know that randn would generate a set Gaussian > > samples. But if I need to add gaussian noise to my > > signal such that the noise power is some value N, > > how do I do it? > > How would the variance (that I would give to my rand > > function) var vary as N? > > Please explain? > > > > Your help is appreciated!!! > > > > Regards > > Rohit Sharma > > > > > > Rohit Sharma, > > 65,Villages at North Gate, > > 3000, July Street > > Baton Rouge, > > LA, 70803 > > USA. > > "We met as total strangers and ended as lifetime > > friends, we shared a season of our lives, but every > > season has to end." > > > > --------------------------------- > > |