I saw this adding noise to a signal but I still have some questions.

I have normalized IQ data of a signal. Lets suppose that I know what the current SNR is (say 8dBm). How can I add noise to the I and Q components of my signal so that it has a new SNR (of course one with a worse SNR then what I am starting off with)?

My problem is 2 fold, the first one is, how can I add a precise amount of noise to an already noisy signal? I want to control what the new SNR will be, so I need to be careful how much noise I add.

The second is, technically, I have normalized I and Q data. How do I add noise to the I and Q components? Or, do I just need to recreate the signal with I and Q, add the noise, and then break it back up into I and Q components?

Thank you all for your help :)


1 Answer 1


UPDATE**** Attempt # 2: After thinking about this overnight, I do not believe that looking at the peaks does anything useful. SNR is a measure of a signal power to signal noise, so I think I need to look at the power of these signals. With that in mind, here is my thought: We start with:

S = clean signal power

N = noise power

Ps = observer S + n power (I can measure this)

SNR = signal to noise ratio in watts

S/N = SNR, with this being a known SNR in watts.

S = SNR*N, re-writing the equation

Ps = S + N

Ps = SNR*N+N = N(1+SNR)

N = Ps/(1+SNR), so now we can know the noise power

S = SNR*N, and we can find the signal power

If the above is correct, then when we go to add a new noise, we can measure the power of this noise signal before adding it, call it N2.

(S)/(N+N2) = newSNR

Does that look correct?

Thank you all for your help.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.