ir_smooth {ir}

Smooths infrared spectra in an ir object

Description

ir_smooth applies smoothing functions to infrared spectra. ir_smooth either performs Savitzky-Golay smoothing, using on signal::sgolayfilt(), or Fourier smoothing using fda::smooth.basis(). Savitzky-Golay smoothing can also be used to compute derivatives of spectra.

Usage

ir_smooth(
  x,
  method = "sg",
  p = 3,
  n = p + 3 - p%%2,
  ts = 1,
  m = 0,
  k = 111,
  ...
)

Arguments

x

An object of class ir.

method

A character value specifying which smoothing method to apply. If method = "sg", a Savitzky-Golay filter will be applied on the spectra. The Savitzky-Golay smoothing will be performed using the function signal::sgolayfilt(). If method = "fourier", Fourier smoothing will be performed. Fourier transformation of the spectra is performed using the fast discrete Fourier transformation (FFT) as implemented in fda::smooth.basis(). A smoothing function can be defined by the argment f.

p

An integer value representing the filter order (i.e. the degree of the polynom) of the Savitzky-Golay filter if method = "sg".

n

An odd integer value representing the length (i.e. the number of wavenumber values used to construct the polynom) of the Savitzky-Golay filter if method = "sg".

ts

time scaling factor. See signal::sgolayfilt().

m

An integer value representing the mth derivative to compute. This option can be used to compute derivatives of spectra. See signal::sgolayfilt().

k

A positive odd integer representing the number of Fourier basis functions to use as smoothed representation of the spectra if method = "fourier".

...

additional arguments (ignored).

Details

When x contains spectra with different wavenumber values, the filters are applied for each spectra only on existing wavenumber values. This means that the filter window (if method == "sg") will be different for these different spectra.

Value

x with smoothed spectra.

Examples

#' # Savitzky-Golay smoothing
x1 <-
   ir::ir_sample_data[1:5, ] %>%
   ir::ir_smooth(method = "sg", p = 3, n = 51, ts = 1, m = 0)

# Fourier smoothing
x2 <-
   ir::ir_sample_data[1:5, ] %>%
   ir::ir_smooth(method = "fourier", k = 21)

# computing derivative spectra with Savitzky-Golay smoothing (here: first
# derivative)
x3 <-
   ir::ir_sample_data[1:5, ] %>%
   ir::ir_smooth(method = "sg", p = 3, n = 51, ts = 1, m = 1)