Low Pass Filter

Use the Database Tools > Filters > Low Pass Filter menu option (LOWPASS GX) to apply a low-pass (smoothing) filter to a channel.

Low Pass Filter dialog options

Channel to filter	Name of the input channel to be filtered Script Parameter: LOWPASS.IN
Output channel	Name of the filtered output channel Script Parameter: LOWPASS.OUT
Cutoff wavelength (fiducials)	Short wavelength cutoff in fiducials. Script Parameter: LOWPASS.CUTOFF

Application Notes

A low-pass filter smoothes the input data by the application of a convolution filter to the data. The filter is called 'low-pass' because it allow slow wavenumbers (low frequencies) to pass to the output channel. Features in the data that are shorter than the short wavelength cutoff will removed. The convolution filter is designed using the method of Fraser, 1966.The default length will be as long as the cutoff wavelength, which produces a filter that exhibits a minimum of side effects, such as Gibb's phenomena.

A low-pass filter is simply a band-pass filter with the long-wavelength cutoff set to a very large number. If you would like to control the length of the filter, use the BANDPASS GX and set the filter length directly.

A low-pass filter is considered linear because all data is treated by the filter equally. This can be a problem when attempting to remove short-wavelength, but high amplitude features because quite strong filters may be required. Such filters can affect parts of the data that you did not intend to disturb. An alternative is to use a non-linear filter such as NLFILT.