dsp::Response Class Reference

Describes a frequency (or impulse) response. More...

#include <Response.h>

Inheritance diagram for dsp::Response:

Public Member Functions
	Response ()
	Default constructor.
virtual	~Response ()
	Destructor.
	Response (const Response &)
	Copy constructor.
const Response &	operator= (const Response &)
	Assignment operator.
const Response &	operator*= (const Response &)
	Multiplication operator.
virtual void	prepare (const Observation *input, unsigned channels=0)
	Calculate the impulse_pos and impulse_neg attributes.
virtual void	match (const Observation *input, unsigned channels=0)
	Match the frequency response to the input Observation. More...
virtual void	match (const Response *response)
	Match the frequency response to another Response.
unsigned	get_input_nchan () const
	Get the number of input channels.
void	set_input_nchan (unsigned _input_nchan)
	Set the number of input channels.
unsigned	get_impulse_pos () const
	Get the size of the positive half of the impulse response, . More...
unsigned	get_impulse_neg () const
	Get the size of the negative half of the impulse response, . More...
void	set_impulse_pos (unsigned _impulse_pos)
	Set the size of the positive half of the impulse response, . More...
void	set_impulse_neg (unsigned _impulse_neg)
	Set the size of the negative half of the impulse response, . More...
virtual void	set_dc_centred (bool dc_centred)
	Set the flag for a bin-centred spectrum.
bool	get_dc_centred () const
	Get the flag for a bin-centred spectrum.
virtual void	set_swap (bool flag)
	Set the flag for a band-swapped spectrum.
bool	get_swap () const
	Get the flag for a band-swapped spectrum.
unsigned	get_minimum_ndat () const
	Return the minimum useable ndat. More...
void	set_optimal_ndat ()
	Resize with ndat set to the optimal value. More...
void	set_optimal_fft (OptimalFFT *)
	Set the policy used to compute the optimal FFT length.
OptimalFFT *	get_optimal_fft () const
bool	has_optimal_fft () const
void	check_ndat () const
	Given impulse_pos and impulse_neg, check that ndat is large enough.
std::vector< float >	get_passband (unsigned ipol, int ichan=-1) const
	Get the passband.
void	set (const std::vector< std::complex< float > > &phasors)
	Construct frequency response from complex phasors.
void	set (const std::vector< Jones< float > > &jones)
	Construct frequency response from jones matrices.
void	operate (float *spectrum, unsigned poln=0, int ichan=-1) const
	Multiply spectrum by complex frequency response. More...
void	operate (float *spectrum, unsigned poln, int ichan_start, unsigned nchan_op) const
	Multiply spectrum by complex frequency response. More...
void	operate (float *input_spectrum, float *output_spectrum, unsigned poln, int ichan_start, unsigned nchan_op) const
	Multiply spectrum by complex frequency response, out-of-place. More...
void	operate (float *spectrum1, float *spectrum2, int ichan=-1) const
	Multiply spectrum vector by complex matrix frequency response.
void	integrate (float *spectrum, unsigned poln=0, int ichan=-1)
	Integrate the power of spectrum into self. More...
void	integrate (float *spectrum1, float *spectrum2, int ichan=-1)
	Integrate coherency matrix of vector spectrum into self.
virtual bool	matches (const Shape *response)
	Returns true if the dimension and ordering match.
virtual void	mark (Observation *output)
	Modify the out Observation information as seen fit by sub-classes.
void	naturalize ()
	Re-organize frequency bins to reflect natural ordering (DC->Nyq)
virtual bool	state_is_valid (std::string &reason)
	Enable Response to be used in Transformation template.
void	flagswap (unsigned divisions=1)
	Set flags that response should be swapped.
float	get_nbyte () const
	Requires zero bytes per time sample.
void	check_finite (const char *name)
	Worker function checks that every value in response is finite.
Public Member Functions inherited from dsp::Shape
	Shape ()
	Default constructor.
virtual	~Shape ()
	Destructor.
	Shape (const Shape &)
	Copy constructor.
const Shape &	operator= (const Shape &)
	Assignment operator.
virtual void	resize (unsigned npol, unsigned nchan, unsigned ndat, unsigned ndim)
	Set the dimensions of the data.
unsigned	get_npol () const
	Get the number of polarizations.
unsigned	get_nchan () const
	Get the number of frequency channels.
unsigned	get_ndat () const
	Get the number of datum in each of the nchan*npol divisions.
unsigned	get_ndim () const
	Get the dimension of each datum (e.g. 2=complex 8=Jones)
void	scrunch_to (unsigned ndat)
	Scrunch each dimension to a new ndat.
void	rotate (int npt)
	Rotate data so that Shape[i] = Shape[i+npt].
void	zero ()
	Set all values to zero.
void	zero_dimensions ()
	Set all dimnensions to zero; allocated memory unaffected.
void	borrow (const Shape &, unsigned ichan=0)
	Borrow the data from the specified channel of another Shape.
const Shape &	operator/= (float factor)
	Divide each point by factor.
const Shape &	operator*= (float factor)
	Multiply each point by factor.
const Shape &	operator+= (const Shape &)
	Add another Shape to this one.
float *	get_datptr (unsigned ichan, unsigned ipol)
	Provide access to the data for the specified polarization.
const float *	get_datptr (unsigned ichan, unsigned ipol) const
	Provide access to the data for the specified polarization.
Public Member Functions inherited from OwnStream
	OwnStream (const OwnStream &)
const OwnStream &	operator= (const OwnStream &)
virtual void	set_cout (std::ostream &) const
virtual void	set_cerr (std::ostream &) const
Public Member Functions inherited from Reference::Able
	Able (const Able &)
Able &	operator= (const Able &)
unsigned	get_reference_count () const
Public Member Functions inherited from Reference::HeapTracked
	HeapTracked (const HeapTracked &)
HeapTracked &	operator= (const HeapTracked &)
bool	__is_on_heap () const

Public Attributes
Callback< Response >	changed
	Called when the Response has been changed.

Static Public Attributes
static unsigned	ndat_max = 0
	Maximum ndat allowed. More...
Static Public Attributes inherited from dsp::Shape
static bool	verbose = false

Protected Member Functions
void	calc_lcf (unsigned a, unsigned b, const Rational &osf, std::vector< unsigned > &result)
void	calc_oversampled_discard_region (unsigned *_discard_neg, unsigned *_discard_pos, unsigned _nchan, const Rational &osf)
void	calc_oversampled_fft_length (unsigned *_fft_length, unsigned _nchan, const Rational &osf, int direction=-1)
void	doswap (unsigned divisions=1)
	Swap halves of bandpass(es)
Protected Member Functions inherited from dsp::Shape
void	init ()
void	size_dataspace ()
void	destroy ()
Protected Member Functions inherited from Reference::Able
Handle *	__reference (bool active) const
void	__dereference (bool auto_delete=true) const

Protected Attributes
unsigned	step
Reference::To< OptimalFFT >	optimal_fft
unsigned	impulse_pos
	Complex time samples in t>0 half of corresponding impulse response.
unsigned	impulse_neg
	Complex time samples in t<0 half of corresponding impulse response.
bool	whole_swapped
	Toggled every time Response::doswap(1) is called (default: false)
unsigned	swap_divisions
	Toggled every time Response::doswap(divisions > 1) is called.
bool	dc_centred
	Toggled when built for a bin-centred spectrum.
unsigned	input_nchan
	number of input channels. Used by some derived classes
Protected Attributes inherited from dsp::Shape
float *	buffer
	Data points.
unsigned	bufsize
	Size of the data buffer.
unsigned	offset
	Offset between datum from each polarization.
unsigned	npol
	Number of polarizations.
unsigned	nchan
	Number of frequency divisions (channels)
unsigned	ndat
	Number of datum in each of the npol*nchan divisions.
unsigned	ndim
	Dimension of each datum.
bool	borrowed
	Flag that data are borrowed from another Shape.

Additional Inherited Members
Static Public Member Functions inherited from Reference::Able
static size_t	get_instance_count ()
Static Public Member Functions inherited from Reference::HeapTracked
static void *	operator new (size_t size, void *ptr=0)
static void	operator delete (void *location, void *ptr)
static void	operator delete (void *location)
static size_t	get_heap_queue_size ()

Detailed Description

Describes a frequency (or impulse) response.

Member Function Documentation

get_impulse_neg()

unsigned dsp::Response::get_impulse_neg ( ) const

inline

Get the size of the negative half of the impulse response, .

Get the number of complex time samples in the t<0 half of the corresponding impulse response function

References impulse_pos.

Referenced by dsp::Archiver::pack().

get_impulse_pos()

unsigned dsp::Response::get_impulse_pos ( ) const

inline

Get the size of the positive half of the impulse response, .

Get the number of complex time samples in the t>0 half of the corresponding impulse response function

References impulse_neg.

Referenced by dsp::Archiver::pack().

get_minimum_ndat()

unsigned dsp::Response::get_minimum_ndat

(

)

const

Return the minimum useable ndat.

Using the impulse_pos and impulse_neg attributes, this method determines the minimum acceptable ndat for use in convolution. This is given by the smallest power of two greater than or equal to the twice the sum of impulse_pos and impulse_neg.

integrate()

void dsp::Response::integrate	(	float *	data,
		unsigned	ipol = 0,
		int	ichan = -1
	)

Integrate the power of spectrum into self.

Adds the square of each complex point to the current power spectrum

Parameters

data	an array of nchan*ndat complex numbers
ipol	the polarization of the data (Response may optionally integrate a different power spectrum for each polarization)

match()

void dsp::Response::match ( const Observation *  input, unsigned  channels = 0  )

virtual

Match the frequency response to the input Observation.

The ordering of frequency channels in the response function depends upon:

• the state of the input Observation (real or complex); and
• Operations to be performed upon the Observation (e.g. simultaneous filterbank)

As well, sub-classes of Response may need to dynamically check, refine, or define their frequency response function based on the state of the input Observation or the number of channels into which it will be divided.

Parameters

input	Observation to which the frequency response is to be matched
channels	If specified, the number of filterbank channels into which the input Observation will be divided. Response::match does not use this parameter, but sub-classes may find it useful.

Reimplemented in dsp::ScalarFilter, dsp::ResponseProduct, and dsp::RFIFilter.

operate() [1/3]

void dsp::Response::operate	(	float *	input_spectrum,
		float *	output_spectrum,
		unsigned	poln,
		int	ichan_start,
		unsigned	nchan_op
	)		const

Multiply spectrum by complex frequency response, out-of-place.

Multiply spectrum by complex frequency response.

operate() [2/3]

void dsp::Response::operate	(	float *	spectrum,
		unsigned	poln,
		int	ichan_start,
		unsigned	nchan_op
	)		const

Multiply spectrum by complex frequency response.

Multiply spectrum by complex frequency response in specified channels.

operate() [3/3]

void dsp::Response::operate	(	float *	data,
		unsigned	ipol = 0,
		int	ichan = -1
	)		const

Multiply spectrum by complex frequency response.

Multiplies an array of complex points by the complex response

Parameters

ipol	the polarization of the data (Response may optionally contain a different frequency response function for each polarization)
data	an array of nchan*ndat complex numbers

Referenced by operator*=().

set_impulse_neg()

void dsp::Response::set_impulse_neg ( unsigned  _impulse_neg )

inline

Set the size of the negative half of the impulse response, .

Set the number of complex time samples in the t<0 half of the corresponding impulse response function

set_impulse_pos()

void dsp::Response::set_impulse_pos ( unsigned  _impulse_pos )

inline

Set the size of the positive half of the impulse response, .

Set the number of complex time samples in the t>0 half of the corresponding impulse response function

References impulse_neg.

set_optimal_ndat()

void dsp::Response::set_optimal_ndat

(

)

Resize with ndat set to the optimal value.

Using the get_minimum_ndat method and the max_ndat static attribute, this method determines the optimal ndat for use in convolution.

Member Data Documentation

ndat_max

unsigned dsp::Response::ndat_max = 0

static

Maximum ndat allowed.

If specified, this attribute restricts the value for ndat chosen by the set_optimal_ndat method, enabling the amount of RAM used by the calling process to be limited.

---

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

• Response.h
• Response.C