I have a question regarding the concept of carrier power, output power and power relative to carrier (dBc). Suppose I have a carrier frequency of 4Hz and signal bandwidth = 4 Hz. The signal span is from [2Hz 6Hz].
My indexing starts at 0 for simplicity.

  1. I have 7 samples with sampling frequency 7 Hz such that:

    x(n)=[1+1j 2-2j 1+3j 3+j 4 5 2+2j]; % Time domain signal

Representing the signal into frequency domain by taking the FFT with 7 points (each frequency bin is 1 Hz range 0-6 Hz) and normalizing for power,



X[n]=[2.6+0.7j -0.56+0.54j -0.99-0.52j -0.5+0.07j 0.38+0.69j 0.7+0.2j -0.63-0.7j].
The output power in dBm within the Bandwidth from 2Hz to 6Hz (for simplicity we are considering up to only 4 samples excluding 6Hz sample) is:

OutPow=10*log10(sum(abs(X[:,2:5]).^2))+30; % in dBm

Therefore, the signal power is 34.34 dBm. Note: I don't think this is the average power.

I would like to generate a noise with -20 dBc. Here are two methods to generate noise power with -20dBc in frequency domain.

Method 1 (the idea is Wanted Signal Power (BW 4 Hz) -Noise at the carrier = 20 dBc):
SigPow=34.34 ; %dBm
TargetNoisePower=(34.34-20)=5.66; % Its down to 20 dBc
NoiseCarrierMagnitude=sqrt(1/1000)*10^(5.66/20)=0.0607; % Noise amplitude for carrier
Noise(n)=[0 0 0 0 0.0607 0 0];  % A single sample at frequency domain

Method 2 (The idea is Sample Power at carrier ($f_c=4$ Hz) -Noise at the carrier =20 dBc):

FcPow=27.93 ; %dBm At 4Hz the sample is 0.38+0.69j
TargetNoisePower=(27.93-20)=7.93;  % at the carrier
NoiseCarrierMagnitude=sqrt(1/1000)*10^(7.93/20)=0.079; % Noise amplitude for carrier
Noise(n)=[0 0 0 0 0.079 0 0];  % A single sample at frequency domain

My questions are the following:

  1. Which method is right for generating a noise signal in dBc?
  2. The channel might not be occupied with signal at the carrier frequency. For example within the signal BW the exact carrier frequency ($f_c=4$ Hz) sample might be 0 or have only some noise (AWGN) sample.
  3. When we say signal power within the bandwidth, do we mean the total power within the bandwidth?
  • $\begingroup$ your signal doesn't seem to have a carrier? $\endgroup$ Aug 25, 2022 at 8:00
  • $\begingroup$ Isn't there a problem with Nyquist as well here? Sampling rate should be greater than 2*BW $\endgroup$
    – Jdip
    Aug 25, 2022 at 11:29
  • 1
    $\begingroup$ @Jdip nope, if we're doing it like this, it's basically direct synth from the DFT; we're in complex baseband! $\endgroup$ Aug 25, 2022 at 12:22
  • $\begingroup$ @MarcusMüller yup, thanks! $\endgroup$
    – Jdip
    Aug 25, 2022 at 14:55

1 Answer 1


Your questions:

  1. You might have a carrier frequency, but you don't seem to have a carrier in your signal, so dBc is not a definable concept. You can only generate a noise with a given power in dB of carrier power if there's a carrier.
  2. Not a question!
  3. Yes


  • Your OutPow is not a power, the way you calculate it, but an energy, it's thus also not in dB of milliwatt.
  • I think what you're really just after is an SNR in dB. That's just the ratio of average signal power to average noise power.
  • $\begingroup$ Muller thank you for an explanation. $\endgroup$ Aug 25, 2022 at 19:49
  • $\begingroup$ However, I think you might be wrong in number 1 because power in frequency domain sum(X(k).^2)/Ndfft^2, where X(k) is the fft sample. If possible please clarify my doubt. (2) Sorry for not articulating my question in a proper way, but "You can only generate a noise with a given power in dB of carrier power if there's a carrier." what do you mean by if there is a carrier. Suppose for some signal set "B" Fc-is the center frequency which might interfere with an existing signal. The power defference between P_Sig-P_B=X; And X is in dBc. Am I missing somethin. $\endgroup$ Aug 25, 2022 at 20:01
  • $\begingroup$ Just because you're centered around some center requency doesn't mean you transmit a carrier. But stating something in dBc requires that there is a carrier. There is really nothing to clarify here: without a carrier you can't state something as relative to the power of the carrier. $\endgroup$ Aug 25, 2022 at 22:19
  • 1
    $\begingroup$ As I already explained in my answer what you describe is not power relative to a carrier, but simply SNR. $\endgroup$ Aug 25, 2022 at 22:20

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.