How to generate AWGN with correct PSD on MATLAB

[@varunsharma-fs8vz]

Thanks for the video, I have a question that fft after reshaping will do a fft on 100 samples or 4096 samples?

[@sohamyadav5755]

hi, I have a doubt with my matlab code , its a eeg signal code in which I've used awng, can u help me with it!

[@krishnakgurumoorthy8879]

Hi Igor, very nice video indeed!. I am able to get the correct psd but I am unable to get proper scaling when I add a signal to the noise voltage. My understanding is Noise powers add and that is why we have a sum(window) to get the proper scaling. So When I add a voltage signal to noise, averaging multiple captures works to lower the noise variance but signal power does not scale properly. I will send an email with my code.

[@igorauadfr]

Thanks for watching. I will try to help you over e-mail.

[@bhimsgk]

Hi, I generated AM signal with a SNR of 5dB and SNR of 10dB , am i suppose to see PSD computed to be the same ? thanks

[@yanjiang8920]

Hello, i wonder why noiseFloor minus 30 is the process of converting dBm/Hz to dB/Hz?

[@igorauadfr]

Thanks for the question. PdBm is by definition 10*log10(Pmw), where Pmw is the power in milliwatts. PdB is 10*log10(Pw), where Pw is the power in watts. Hence, PdBm = 10*log10(Pmw) = 10*log10(10^3 * Pw) = 10*log10(10^3) + 10*log10(Pw) = 30 + 10*log10(Pw) = 30 + PdB Therefore, PdB = PdBm - 30

[@yanjiang8920]

Igor Freire Thank you very much. It is very helpful.

[@rajvirsingh6079]

Hi, nice video - thanks for posting. Just a question about the final plot. Your x-axis went from 0 to 200 MHz. The bandwidth was defined as 100 MHz, and then the sampling frequency 200 MHz. Should that x-axis be -100 to 100 MHz?

[@igorauadfr]

rajvirsingh Good question. What goes between 100MHz and 200MHz is the same as what would be within -100MHz to 0. It is quite common to adopt the frequency axis starting from 0 up to the sampling frequency fs (or from 0 to 2*pi), instead of starting from -fs/2 up to fs/2 (or -pi to pi). If you want to plot the spectrum from -fs/2 to fs/2, you can use the Matlab function "fftshift()".

[@rajvirsingh6079]

Igor Freire Thanks for the reply!