R/Sinc_interp.R

In wltcwpf/RCTC: The Rotated Combination of Two-Component Ground Motions

#' A Sinc interpolation Function
#'
#' This function applies Sinc interpolation to reduce the time step and
#' returns a time series with more data points than in \code{data}
#' @param data An acceleration time series array
#' @param interpolation_factor An interpolated factor, specifies the degree of interpolation
#' and should be a power of two.
#' @return The interpolated time series of \code{data} by factor \code{interpolation_factor}
#' @keywords internal
#' @importFrom pracma zeros ifft
#' @importFrom stats fft
#' @export

########################################
######## inputs: time series, interpolation multiplier
######## outputs: Sinc interpolated time series
Interpft <- function(data, interpolation_factor){
  if (interpolation_factor == 1) {
    return(data)
  } else {
    npts <- length(data)
    npw2_exp = floor(log(npts)/log(2))
    npw2_temp = 2^npw2_exp
    if(npw2_temp == npts){
      npw2 = npts
    } else{
      npw2 <- 2^(npw2_exp+1)
      data <- c(data, pracma::zeros((npw2-npts),1))
    }
    ## ensure length(data) is a power of 2
    nyqst = npw2/2 + 1
    m = length(data)
    a = stats::fft(data)
    b = c(a[1:nyqst], pracma::zeros((m*interpolation_factor-m),1), a[(nyqst+1):m])
    b[nyqst] = b[nyqst]/2
    b[nyqst+m*interpolation_factor-m] = b[nyqst]
    y = pracma::ifft(b)
    y = y*interpolation_factor
    y = Re(y)
    y <- y[1:(npts*interpolation_factor)]
  }
}