Skip to main content

Short note: White noise and its PSDF

Short answer

White noise has a constant power spectral density S_n(f)=N_0/2 for all f, and its autocorrelation is(N_0/2)δ(τ). It is an idealised model; thermal noise approximates it up to ~10^12 Hz, behaving as white within practical bandwidths.

Write a short note on white noise and its psdf.

White noise is an idealised random noise whose power spectral density is constant at all frequencies, just as white light contains all colours equally. Its two-sided PSD is

Sn(f)=N02for all fS_n(f) = \frac{N_0}{2} \quad \text{for all } f

Its autocorrelation is the inverse Fourier transform, an impulse:

Rn(τ)=N02 δ(τ)R_n(\tau) = \frac{N_0}{2}\,\delta(\tau)

so any two different samples are uncorrelated (and independent if Gaussian).

 Sn(f)                      Rn(tau)
   | N0/2                      ^ (N0/2) delta
 --+---------------- f   ------+------ tau
  • Total power ∫Sndf\int S_n df is infinite, so it is only a model; real noise (thermal, shot) is flat up to very high frequencies (∼1012\sim 10^{12} Hz), so it acts white in any practical bandwidth.
  • For thermal noise N0=kTN_0 = kT. Noise through a filter of bandwidth BB has power N0BN_0B.
  • Usually also assumed Gaussian: AWGN.

✓ Answer checked by Claude Opus 5.5 (high effort)

All Communication Systems chapter 9 questions →