signal_filter function

Filter a seismic signal in the time or frequency domain

The function filters the input signal vector in the time or frequency domain.

signal_filter(
  data,
  f,
  fft = FALSE,
  dt,
  type,
  shape = "butter",
  order = 2,
  p,
  zero = FALSE,
  lazy = FALSE
)

Arguments

• data: eseis object, numeric vector or list of objects, data set to be processed.

• f: Numeric value or vector of length two, lower and/or upper cutoff frequencies (Hz).

• fft: Logical value, option to filter in the time domain (fft = FALSE) or the frequency domain (fft = TRUE). Default is (fft = FALSE).

• dt: Numeric value, sampling rate If omitted, dt

  is set to 1/200.

• type: Character value, type of filter, one out of "LP" (low pass), "HP" (high pass), "BP" (band pass) and "BR" (band rejection). If omitted, the type is interpreted from f. If f is of length two, type is set to "BP". If f is of length one, type is set to "HP".

• shape: Character value, one out of "butter"

  (Butterworth), default is "butter".

• order: Numeric value, order of the filter, default is 2. Only needed if data is no eseis object.

• p: Numeric value, fraction of the signal to be tapered. If omitted, no tapering will be done.

• zero: Logical value, option to run filter in zero phase shift mode. Note that this will triple the length of the signal vector during calculation. Default is FALSE.

• lazy: Logical value, option to pre- and post-process data, including detrending, demeaning and tapering (p = 0.02). Default if FALSE.

Returns

Numeric vector or list of vectors, filtered signal vector.

Examples

## load example data set
data(rockfall)

## filter data set by bandpass filter between 1 and 90 Hz
rockfall_bp <- signal_filter(data = rockfall_eseis, 
                             f = c(1, 90))
                             
## taper signal to account for edge effects
rockfall_bp <- signal_taper(data = rockfall_bp, 
                            n = 2000)

## plot filtered signal
plot_signal(data = rockfall_bp)

## compare time domain versus frequency domain filtering
rockfall_td <- signal_filter(data = rockfall_eseis, 
                             f = c(10, 40), 
                             fft = FALSE)
                             
rockfall_td_sp <- signal_spectrum(data = rockfall_td)

rockfall_fd <- signal_filter(data = rockfall_eseis, 
                             f = c(10, 40), 
                             fft = TRUE)
                             
rockfall_fd_sp <- signal_spectrum(data = rockfall_fd)

plot_spectrum(data = rockfall_td_sp)
plot_spectrum(data = rockfall_fd_sp)

Author(s)

Michael Dietze