noiselevel

Estimate the noise standard deviation on a signal.

Syntax

sigma = noiselevel(V2D)
sigma = noiselevel(Vc)
sigma = noiselevel(V)
sigma = noiselevel(V,filter)
sigma = noiselevel(V,Vref)
Parameters
  • V2D - 2D-dataset of scans of a dipolar signal (NxM-element matrix)
  • Vc - 1D complex valued dipolar signal (N-element array)
  • V - 1D dipolar signal (N-element array)
  • filter - Filtering method (string)
  • Vref - Reference signal (N-element array)
Returns
  • sigma - Estimated noise standard deviation (scalar)

Description

sigma = noiselevel(V2D)

If a 2D-dataset V2D of different scans for a signal is provided, the noise standard deviation is estimated from the deviations between scans. The second dimension of V2D must contain the different scans. The function returns the standard deviation of the averaged signal not of the individual scans.

sigma = noiselevel(Vc)

If a complex-valued signal Vc is provided, the imaginary component is minimzed via phase correction and the noise standard deviation is estimated from the phase-corrected imaginary component.

sigma = noiselevel(V)
sigma = noiselevel(V,filter)

If a real-valued signal V is provided, the noise standard deviation is estimated from the deviation obtained via application of a moving mean filter. The nature of the filter can specified by setting filter to one of the following methods:

  • 'movmean' - Moving mean filter (default)
  • 'savgol' - Savitzky-Golay filter
sigma = noiselevel(V,Vref)

If a reference model signal Vref is given, the noise level is estimated from the difference between both signals.