#include <Filters.h>

Static Public Member Functions
static void	Bilateral2Filter (double *&, int, int, double, double, double)
	Nonlinear filter based on bilateral filter.

static void	BilateralFilter (double *&, int, double, double, double)
	Bilateral filter (non-linear) with 3 sigma kernel.

static void	BoxFilter (double *&, int, int)
	Simple running average (box) filter.

static void	Convolute (double &, double , double *, int)
	Helper to perform convolution of two 1D arrays.

static void	Convolute (double &, double , int)
	Helper to perform convolution of two 1D arrays.

static void	Deconvolute (double &, double , double *, int, double, double=0., double=0.)
	Helper to perform partial deconvolution of two 1D arrays.

static void	GausFFTFilter (double *&, int, double, double)
	Gaussian smoothing with FFT convolution.

static void	GausFilter (double *&, int, double, double)
	Gaussian smoothing with simple 3 sigma kernel.

static void	MedianFilter (double *&ar, int, int)
	Median filter.

static void	ResponseFilter (double *&, int, double=.4, double=1.2, double=.25, double=.3125)
	Custom filter emulating primitive response function.

static void	SecondOrderUnderdampedFilter (double *&, int, double, double, double, double=.3125, bool=false)
	Shifted second order underdamped filter.

static void	SecondOrderUnderdampedFilter (TH1F *&, int, double, double, double, double=.3125, bool=false)
	Shifted second order underdamped filter See above for details.

static void	SmoothArray (double *&, int, double, int, double=.3125, double=1.5)
	Apply smoothing array of double with length nbins.

static void	SmoothArray (double *&, int, double=.625, string="Gaus", double=.3125, double=1.5)
	Apply smoothing array of double with length nbins.

Detailed Description

Definition at line 21 of file Filters.h.

Member Function Documentation

◆ Bilateral2Filter()

void Filters::Bilateral2Filter	(	double *&	ar,
		int	nbins,
		int	sigma_x,
		double	sigma_y,
		double	sigma_slope,
		double	bin_size
	)

static

Nonlinear filter based on bilateral filter.

Uses data and slope of data to weight the smoothing. x is not used. Very slow.

Parameters

ar	Array to be smoothed.
nbins	Number of bins of input.
sigma_x	Window along x in bins.
sigma_y	Sigma along y in mV.
sigma_slope	Sigma of slope in mV.
bin_size	Bin size in ns.

Definition at line 338 of file Filters.cc.

◆ BilateralFilter()

void Filters::BilateralFilter	(	double *&	ar,
		int	nbins,
		double	sigma_x,
		double	sigma_y,
		double	bin_size
	)

static

Bilateral filter (non-linear) with 3 sigma kernel.

For noise reduction while preserving edges, i .e. for baseline correction. Slow.

Parameters

ar	Array to be smoothed.
nbins	Number of bins of input.
sigma_x	Gauss sigma along x in ns.
sigma_y	Gauss sigma along y in mV.
bin_size	Bin size in ns.

Definition at line 299 of file Filters.cc.

◆ BoxFilter()

void Filters::BoxFilter	(	double *&	ar,
		int	nbins,
		int	sigma
	)

static

Simple running average (box) filter.

Parameters

ar	Array to be smoothed.
nbins	Number of bins of input.
sigma	Number of bins before and after central bin.

Definition at line 207 of file Filters.cc.

◆ Convolute() [1/2]

void Filters::Convolute	(	double *&	result,
		double *	first,
		double *	second,
		int	size
	)

static

Helper to perform convolution of two 1D arrays.

WARNING: All arrays must be of the same size.

Used for smoothing etc.

Parameters

[in,out]	result	Array containing convolution result
	first	First array for convolution
	second	Second array for convolution
	size	Size of arrays

Definition at line 12 of file Filters.cc.

◆ Convolute() [2/2]

void Filters::Convolute	(	double *&	first,
		double *	second,
		int	size
	)

static

Helper to perform convolution of two 1D arrays.

WARNING: All arrays must be of the same size.

Used for smoothing etc.

Parameters

[in,out]	first	Array to be convoluted
	second	Second array for convolution
	size	Size of arrays

Definition at line 65 of file Filters.cc.

◆ Deconvolute()

void Filters::Deconvolute	(	double *&	result,
		double *	first,
		double *	second,
		int	size,
		double	sigma,
		double	x0 = `0.`,
		double	bin_size = `0.`
	)

static

Helper to perform partial deconvolution of two 1D arrays.

WARNING: All arrays must be of the same size.

Parameters

[in,out]	result	Array containing deconvolution result
	first	First array for deconvolution
	second	Second array for deconvolution
	size	Size of arrays
	sigma	Gaussian sigma in ns
	x0	Gaussian center in ns
	bin_size	Bin size in ns

Definition at line 80 of file Filters.cc.

◆ GausFFTFilter()

void Filters::GausFFTFilter	(	double *&	ar,
		int	nbins,
		double	sigma,
		double	bin_size
	)

static

Gaussian smoothing with FFT convolution.

Parameters

ar	Array to be smoothed.
nbins	Number of bins of input.
sigma	Gauss sigma in ns.
bin_size	Bin size in ns.

Definition at line 266 of file Filters.cc.

◆ GausFilter()

void Filters::GausFilter	(	double *&	ar,
		int	nbins,
		double	sigma,
		double	bin_size
	)

static

Gaussian smoothing with simple 3 sigma kernel.

Parameters

ar	Array to be smoothed.
nbins	Number of bins of input.
sigma	Gauss sigma in ns.
bin_size	Bin size in ns.

Definition at line 230 of file Filters.cc.

◆ MedianFilter()

void Filters::MedianFilter	(	double *&	ar,
		int	nbins,
		int	window_size
	)

static

Median filter.

Parameters

ar	Array to be smoothed.
nbins	Number of bins of input.
window_size	Window size for median in bins.

Definition at line 378 of file Filters.cc.

◆ ResponseFilter()

void Filters::ResponseFilter	(	double *&	ar,
		int	nbins,
		double	sigma1 = `.4`,
		double	sigma2 = `1.2`,
		double	factor = `.25`,
		double	bin_size = `.3125`
	)

static

Custom filter emulating primitive response function.

Can be used to highlight peaks and suppress long tails (see pole-zero cancellation). Emulates response function of an amplifier setup.
Uses constant padding to avoid edge effects.
Use Filter_test.ipynb to test parameters.

Parameters

[in,out]	ar	Array to be filtered.
	nbins	Number of bins of input.
	sigma1	First.
	sigma2	Second.
	factor	Factor for negative part (<=1).
	bin_size	Bin width. Default is .3125. Set to 1 to get sigma in units of bins.

Definition at line 421 of file Filters.cc.

◆ SecondOrderUnderdampedFilter() [1/2]

void Filters::SecondOrderUnderdampedFilter	(	double *&	ar,
		int	nbins,
		double	period,
		double	damping,
		double	shift,
		double	bin_size = `.3125`,
		bool	do_plot = `false`
	)

static

Shifted second order underdamped filter.

Emulating 2nd order underdamped response (simple damped harmonic oscillator).
Will only consider 3 periods of the response function.
In development, for testing only.

Parameters

[in,out]	ar	Array to be filtered.
	nbins	Number of bins of input.
	period	Period of response function in ns.
	damping	Damping time constant of the response function in ns. Needs to be > 0.
	shift	Shift of response function in ns.
	bin_size	Bin width of the input.
	do_plot	Plot response function and save to file.

Definition at line 466 of file Filters.cc.

◆ SecondOrderUnderdampedFilter() [2/2]

void Filters::SecondOrderUnderdampedFilter	(	TH1F *&	his,
		int	nbins,
		double	period,
		double	damping,
		double	shift,
		double	bin_size = `.3125`,
		bool	do_plot = `false`
	)

static

Shifted second order underdamped filter See above for details.

Parameters

[in,out] his Histogram to be filtered.

Definition at line 547 of file Filters.cc.

◆ SmoothArray() [1/2]

void Filters::SmoothArray	(	double *&	ar,
		int	nbins,
		double	sigma,
		int	method,
		double	bin_size = `.3125`,
		double	sigma2 = `1.5`
	)

static

Apply smoothing array of double with length nbins.

Parameters

method If 0: Use running average (box kernel smoothing). Simple, very fast.
If 1: Use 5 sigma FFT gaussian smoothing. This method is not central and will shift peaks. Very slow.
Else: Use 3 sigma gaussian kernel smoothing. Preferred method, fast.

Definition at line 179 of file Filters.cc.

◆ SmoothArray() [2/2]

void Filters::SmoothArray	(	double *&	ar,
		int	nbins,
		double	sigma = `.625`,
		string	method = `"Gaus"`,
		double	bin_size = `.3125`,
		double	sigma2 = `1.5`
	)

static

Apply smoothing array of double with length nbins.

Use with care. Method="Gaus" is preferred.

Please note that if you want to use gaussian smoothing for data with a binning different from 0.3125 ns/bin you need to set the variable bin_size to the new bin size.

Gaussian smoothing of a simple array with 15 entries. Code in example.

Parameters

[in,out]	ar	Array to be smoothed.
	nbins	Number of bins of input.
	sigma	Number of bins before and after central bin for "Box" and "Median". Gaussian sigma in ns for "Gaus" and "GausFFT". Sigma in x for "Bilateral" and "Bilateral2".
	method	"Box": Use running average (box kernel smoothing). Simple, very fast. "GausFFT": Use 5 sigma FFT gaussian smoothing. This method is not central and will shift peaks. Very slow. "Gaus": Use 3 sigma gaussian kernel smoothing. Preferred method, fast. "Median": Use median filter. Fast. "Bilateral": Use bilateral filter. Slow. "Bilateral2": Use bilateral filter with different kernel. Slow.
	bin_size	Bin width of the array to smooth for gauss sigma. Default is .3125 for wavecatcher sampling rate. Set to 1 to change sigma unit to number of bins.
	sigma2	Sigma in y for "Bilateral" and "Bilateral2".

Examples: use_functions_wo_measurement.cc.

Definition at line 160 of file Filters.cc.

The documentation for this class was generated from the following files:

src/utils/Filters.h
src/utils/Filters.cc

Static Public Member Functions

Detailed Description

Member Function Documentation

◆ Bilateral2Filter()

◆ BilateralFilter()

◆ BoxFilter()

◆ Convolute() [1/2]

◆ Convolute() [2/2]

◆ Deconvolute()

◆ GausFFTFilter()

◆ GausFilter()

◆ MedianFilter()

◆ ResponseFilter()

◆ SecondOrderUnderdampedFilter() [1/2]

◆ SecondOrderUnderdampedFilter() [2/2]

◆ SmoothArray() [1/2]

◆ SmoothArray() [2/2]