Vector and matrix processing toolbox

The vector and matrix processing toolbox consists of a number of classes defined in the namespace MHASignal, and many functions and operators for use with the structures mha_wave_t and mha_spec_t. More...

Collaboration diagram for Vector and matrix processing toolbox:

Classes
struct	mha_wave_t
	Waveform signal structure. More...
struct	mha_audio_descriptor_t
	Description of an audio fragment (planned as a replacement of mhaconfig_t). More...
struct	mha_audio_t
	An audio fragment in the openMHA (planned as a replacement of mha_wave_t and mha_spec_t). More...
class	MHASignal::spectrum_t
	a signal processing class for spectral data (based on mha_spec_t) More...
class	MHASignal::waveform_t
	signal processing class for waveform data (based on mha_wave_t) More...
class	MHASignal::doublebuffer_t
	Double-buffering class. More...
class	MHASignal::hilbert_t
	Hilbert transformation of a waveform segment. More...
class	MHASignal::minphase_t
	Minimal phase function. More...
class	MHASignal::uint_vector_t
	Vector of unsigned values, used for size and index description of n-dimensional matrixes. More...
class	MHASignal::matrix_t
	n-dimensional matrix with real or complex floating point values. More...
class	MHAParser::window_t
	MHA configuration interface for a window function generator. More...
class	MHASignal::delay_wave_t
	Delayline containing wave fragments. More...
class	MHASignal::async_rmslevel_t
	Class for asynchronous level metering. More...

Typedefs
typedef float	mha_real_t
	openMHA type for real numbers More...

Functions
mha_wave_t	range (mha_wave_t s, unsigned int k0, unsigned int len)
	Return a time interval from a waveform chunk. More...
mha_spec_t	channels (mha_spec_t s, unsigned int ch_start, unsigned int nch)
	Return a channel interval from a spectrum. More...
mha_real_t	MHASignal::bin2freq (mha_real_t bin, unsigned fftlen, mha_real_t srate)
	conversion from fft bin index to frequency More...
mha_real_t	MHASignal::freq2bin (mha_real_t freq, unsigned fftlen, mha_real_t srate)
	conversion from frequency to fft bin index More...
mha_real_t	MHASignal::smp2rad (mha_real_t samples, unsigned bin, unsigned fftlen)
	conversion from delay in samples to phase shift More...
mha_real_t	MHASignal::rad2smp (mha_real_t phase_shift, unsigned bin, unsigned fftlen)
	conversion from phase shift to delay in samples More...
template<class elem_type >
std::vector< elem_type >	MHASignal::dupvec (std::vector< elem_type > vec, unsigned n)
	Duplicate last vector element to match desired size. More...
template<class elem_type >
std::vector< elem_type >	MHASignal::dupvec_chk (std::vector< elem_type > vec, unsigned n)
	Duplicate last vector element to match desired size, check for dimension. More...
bool	equal_dim (const mha_wave_t &a, const mha_wave_t &b)
	Test for equal dimension of waveform structures. More...
bool	equal_dim (const mha_wave_t &a, const mhaconfig_t &b)
	Test for match of waveform dimension with mhaconfig structure. More...
bool	equal_dim (const mha_spec_t &a, const mha_spec_t &b)
	Test for equal dimension of spectrum structures. More...
bool	equal_dim (const mha_spec_t &a, const mhaconfig_t &b)
	Test for match of spectrum dimension with mhaconfig structure. More...
bool	equal_dim (const mha_wave_t &a, const mha_spec_t &b)
	Test for equal dimension of waveform/spectrum structures. More...
bool	equal_dim (const mha_spec_t &a, const mha_wave_t &b)
	Test for equal dimension of waveform/spectrum structures. More...
void	integrate (mha_wave_t &s)
	Numeric integration of a signal vector (real values) More...
void	integrate (mha_spec_t &s)
	Numeric integration of a signal vector (complex values) More...
unsigned int	size (const mha_wave_t &s)
	Return size of a waveform structure. More...
unsigned int	size (const mha_spec_t &s)
	Return size of a spectrum structure. More...
unsigned int	size (const mha_wave_t *s)
	Return size of a waveform structure. More...
unsigned int	size (const mha_spec_t *s)
	Return size of a spectrum structure. More...
void	clear (mha_wave_t &s)
	Set all values of waveform to zero. More...
void	clear (mha_wave_t *s)
	Set all values of waveform to zero. More...
void	clear (mha_spec_t &s)
	Set all values of spectrum to zero. More...
void	clear (mha_spec_t *s)
	Set all values of spectrum to zero. More...
void	assign (mha_wave_t self, mha_real_t val)
	Set all values of waveform 'self' to 'val'. More...
void	assign (mha_wave_t self, const mha_wave_t &val)
	Set all values of waveform 'self' to 'val'. More...
void	assign (mha_spec_t self, const mha_spec_t &val)
	Set all values of spectrum 'self' to 'val'. More...
void	timeshift (mha_wave_t &self, int shift)
	Time shift of waveform chunk. More...
mha_real_t &	value (mha_wave_t *s, unsigned int fr, unsigned int ch)
	Access an element of a waveform structure. More...
const mha_real_t &	value (const mha_wave_t *s, unsigned int fr, unsigned int ch)
	Constant access to an element of a waveform structure. More...
mha_complex_t &	value (mha_spec_t *s, unsigned int fr, unsigned int ch)
	Access to an element of a spectrum. More...
const mha_complex_t &	value (const mha_spec_t *s, unsigned int fr, unsigned int ch)
	Constant access to an element of a spectrum. More...
mha_real_t &	value (mha_wave_t &s, unsigned int fr, unsigned int ch)
	Access to an element of a waveform structure. More...
const mha_real_t &	value (const mha_wave_t &s, unsigned int fr, unsigned int ch)
	Constant access to an element of a waveform structure. More...
mha_complex_t &	value (mha_spec_t &s, unsigned int fr, unsigned int ch)
	Access to an element of a spectrum. More...
const mha_complex_t &	value (const mha_spec_t &s, unsigned int fr, unsigned int ch)
	Constant access to an element of a spectrum. More...
std::vector< float >	std_vector_float (const mha_wave_t &)
	Converts a mha_wave_t structure into a std::vector<float> (interleaved order). More...
std::vector< std::vector< float > >	std_vector_vector_float (const mha_wave_t &)
	Converts a mha_wave_t structure into a std::vector< std::vector<float> > (outer vector represents channels). More...
std::vector< std::vector< mha_complex_t > >	std_vector_vector_complex (const mha_spec_t &)
	Converts a mha_spec_t structure into a std::vector< std::vector<mha_complex_t> > (outer vector represents channels). More...
mha_wave_t &	operator+= (mha_wave_t &, const mha_real_t &)
	Addition operator. More...
mha_wave_t &	operator+= (mha_wave_t &, const mha_wave_t &)
	Addition operator. More...
mha_wave_t &	operator-= (mha_wave_t &, const mha_wave_t &)
	Subtraction operator. More...
mha_spec_t &	operator-= (mha_spec_t &, const mha_spec_t &)
	Subtraction operator. More...
mha_wave_t &	operator*= (mha_wave_t &, const mha_real_t &)
	Element-wise multiplication operator. More...
mha_wave_t &	operator*= (mha_wave_t &, const mha_wave_t &)
	Element-wise multiplication operator. More...
mha_spec_t &	operator*= (mha_spec_t &, const mha_real_t &)
	Element-wise multiplication operator. More...
mha_spec_t &	operator*= (mha_spec_t &, const mha_wave_t &)
	Element-wise multiplication operator. More...
mha_spec_t &	operator*= (mha_spec_t &, const mha_spec_t &)
	Element-wise multiplication operator. More...
mha_spec_t &	operator/= (mha_spec_t &, const mha_spec_t &)
	Element-wise division operator. More...
mha_wave_t &	operator/= (mha_wave_t &, const mha_wave_t &)
	Element-wise division operator. More...
mha_spec_t &	operator+= (mha_spec_t &, const mha_spec_t &)
	Addition operator. More...
mha_spec_t &	operator+= (mha_spec_t &, const mha_real_t &)
	Addition operator. More...
mha_wave_t &	operator^= (mha_wave_t &self, const mha_real_t &arg)
	Exponent operator. More...
void	MHASignal::copy_channel (mha_spec_t &self, const mha_spec_t &src, unsigned sch, unsigned dch)
	Copy one channel of a source signal. More...
void	MHASignal::copy_channel (mha_wave_t &self, const mha_wave_t &src, unsigned src_channel, unsigned dest_channel)
	Copy one channel of a source signal. More...
mha_real_t	MHASignal::rmslevel (const mha_spec_t &s, unsigned int channel, unsigned int fftlen)
	Return RMS level of a spectrum channel. More...
mha_real_t	MHASignal::colored_intensity (const mha_spec_t &s, unsigned int channel, unsigned int fftlen, mha_real_t *sqfreq_response=nullptr)
	Colored spectrum intensity. More...
mha_real_t	MHASignal::maxabs (const mha_spec_t &s, unsigned int channel)
	Find maximal absolute value. More...
mha_real_t	MHASignal::rmslevel (const mha_wave_t &s, unsigned int channel)
	Return RMS level of a waveform channel. More...
mha_real_t	MHASignal::maxabs (const mha_wave_t &s, unsigned int channel)
	Find maximal absolute value. More...
mha_real_t	MHASignal::maxabs (const mha_wave_t &s)
	Find maximal absolute value. More...
mha_real_t	MHASignal::max (const mha_wave_t &s)
	Find maximal value. More...
mha_real_t	MHASignal::min (const mha_wave_t &s)
	Find minimal value. More...
mha_real_t	MHASignal::sumsqr_channel (const mha_wave_t &s, unsigned int channel)
	Calculate sum of squared values in one channel. More...
mha_real_t	MHASignal::sumsqr_frame (const mha_wave_t &s, unsigned int frame)
	Calculate sum over all channels of squared values. More...
void	conjugate (mha_spec_t &self)
	Replace (!) the value of this mha_spec_t with its conjugate. More...

Detailed Description

The vector and matrix processing toolbox consists of a number of classes defined in the namespace MHASignal, and many functions and operators for use with the structures mha_wave_t and mha_spec_t.

Typedef Documentation

mha_real_t

typedef float mha_real_t

openMHA type for real numbers

This type is expected to be allways the C-type 'float' (IEEE 754 single).

Function Documentation

range()

mha_wave_t range	(	mha_wave_t	s,
		unsigned int	k0,
		unsigned int	len
	)

Return a time interval from a waveform chunk.

A waveform chunk containing a time intervall of a larger waveform chunk is returned. The number of channels remains constant. The data of the output waveform structure points to the data of the input structure, i.e., write access to the output waveform chunk modifies the corresponding entries in the input chunk.

Parameters

s	Waveform structure
k0	Index of first value in output
len	Number of frames in output

Returns

Waveform structure representing the sub-interval.

channels()

mha_spec_t channels	(	mha_spec_t	s,
		unsigned int	ch_start,
		unsigned int	nch
	)

Return a channel interval from a spectrum.

Parameters

s	Input spectrum
ch_start	Index of first channel in output
nch	Number of channels in output

Returns

Spectrum structure representing the sub-interval.

bin2freq()

mha_real_t MHASignal::bin2freq ( mha_real_t  bin, unsigned  fftlen, mha_real_t  srate  )

inline

conversion from fft bin index to frequency

Parameters

bin	index of fft bin, index 0 has dc
fftlen	FFT length
srate	sampling frequency / Hz

Returns

frequency of fft bin / Hz

freq2bin()

mha_real_t MHASignal::freq2bin ( mha_real_t  freq, unsigned  fftlen, mha_real_t  srate  )

inline

conversion from frequency to fft bin index

Parameters

freq	frequency / Hz
fftlen	FFT length
srate	sampling frequency / Hz

Returns

0-based index of fft bin, generally has non-zero fractional part

smp2rad()

mha_real_t MHASignal::smp2rad ( mha_real_t  samples, unsigned  bin, unsigned  fftlen  )

inline

conversion from delay in samples to phase shift

Compute phase shift that needs to be applied to fft spectrum to achieve the desired delay.

Parameters

samples	delay in samples. Positive delay: shift current signal to future.
bin	index of fft bin, index 0 has dc (index 0 and nyqvist bin cannot be delayed)
fftlen	FFT length

Returns

The phase shift in radiant that needs to be applied to fft bin to achieve the desired delay. A positive delay requires a negative phase shift. If required phase shift is >pi or <-pi, then the desired delay cannot be applied in the fft domain with given parameters. Required phase shifts close to pi should not be used. If bin is 0 or nyqvist, returns 0 phase shift.

rad2smp()

mha_real_t MHASignal::rad2smp ( mha_real_t  phase_shift, unsigned  bin, unsigned  fftlen  )

inline

conversion from phase shift to delay in samples

Compute delay in samples that is achieved by a phase shift.

Parameters

phase_shift	phase shift in radiant
bin	index of fft bin, index 0 has dc (index 0 and nyqvist bin cannot be delayed)
fftlen	FFT length

Returns

The delay in samples achieved by applying the phase shift. A negative phase shift causes a positive delay: shifts current signal to future.

dupvec()

template<class elem_type >

std::vector<elem_type> MHASignal::dupvec	(	std::vector< elem_type >	vec,
		unsigned	n
	)

Duplicate last vector element to match desired size.

Parameters

vec	Input vector.
n	Target number of elements.

Return values

Resized

vector.

dupvec_chk()

template<class elem_type >

std::vector<elem_type> MHASignal::dupvec_chk	(	std::vector< elem_type >	vec,
		unsigned	n
	)

Duplicate last vector element to match desired size, check for dimension.

The input dimension can be either 1 or the target length.

Parameters

vec	Input vector.
n	Target number of elements.

Return values

Resized

vector.

equal_dim() [1/6]

bool equal_dim ( const mha_wave_t &  a, const mha_wave_t &  b  )

inline

Test for equal dimension of waveform structures.

equal_dim() [2/6]

bool equal_dim ( const mha_wave_t &  a, const mhaconfig_t &  b  )

inline

Test for match of waveform dimension with mhaconfig structure.

equal_dim() [3/6]

bool equal_dim ( const mha_spec_t &  a, const mha_spec_t &  b  )

inline

Test for equal dimension of spectrum structures.

equal_dim() [4/6]

bool equal_dim ( const mha_spec_t &  a, const mhaconfig_t &  b  )

inline

Test for match of spectrum dimension with mhaconfig structure.

equal_dim() [5/6]

bool equal_dim ( const mha_wave_t &  a, const mha_spec_t &  b  )

inline

Test for equal dimension of waveform/spectrum structures.

Warning

Waveform structures mha_wave_t use interleaved data order, while spectrum structures mha_spec_t use non-interleaved.

equal_dim() [6/6]

bool equal_dim ( const mha_spec_t &  a, const mha_wave_t &  b  )

inline

Test for equal dimension of waveform/spectrum structures.

Warning

Waveform structures mha_wave_t use interleaved data order, while spectrum structures mha_spec_t use non-interleaved.

integrate() [1/2]

void integrate

(

mha_wave_t &

s

)

Numeric integration of a signal vector (real values)

Parameters

s	Input signal vector

integrate() [2/2]

void integrate

(

mha_spec_t &

s

)

Numeric integration of a signal vector (complex values)

Parameters

s	Input signal vector

size() [1/4]

unsigned int size ( const mha_wave_t &  s )

inline

Return size of a waveform structure.

size() [2/4]

unsigned int size ( const mha_spec_t &  s )

inline

Return size of a spectrum structure.

size() [3/4]

unsigned int size ( const mha_wave_t *  s )

inline

Return size of a waveform structure.

size() [4/4]

unsigned int size ( const mha_spec_t *  s )

inline

Return size of a spectrum structure.

clear() [1/4]

void clear ( mha_wave_t &  s )

inline

Set all values of waveform to zero.

clear() [2/4]

void clear ( mha_wave_t *  s )

inline

Set all values of waveform to zero.

clear() [3/4]

void clear ( mha_spec_t &  s )

inline

Set all values of spectrum to zero.

clear() [4/4]

void clear ( mha_spec_t *  s )

inline

Set all values of spectrum to zero.

assign() [1/3]

void assign ( mha_wave_t  self, mha_real_t  val  )

inline

Set all values of waveform 'self' to 'val'.

Parameters

self	Waveform to be modified.
val	Value to be assigned to all entries of waveform.

assign() [2/3]

void assign	(	mha_wave_t	self,
		const mha_wave_t &	val
	)

Set all values of waveform 'self' to 'val'.

Parameters

self	Waveform to be modified.
val	Source waveform structure.

assign() [3/3]

void assign	(	mha_spec_t	self,
		const mha_spec_t &	val
	)

Set all values of spectrum 'self' to 'val'.

Parameters

self	Spectrum to be modified.
val	Source spectrum.

timeshift()

void timeshift	(	mha_wave_t &	self,
		int	shift
	)

Time shift of waveform chunk.

Shifted areas are filled with zeros.

Parameters

self	Waveform chunk to be shifted
shift	Shift amount, positive values shift to later times

value() [1/8]

mha_real_t& value ( mha_wave_t *  s, unsigned int  fr, unsigned int  ch  )

inline

Access an element of a waveform structure.

Parameters

s	Waveform structure
fr	Frame number
ch	Channel number

Returns

Reference to element

value() [2/8]

const mha_real_t& value ( const mha_wave_t *  s, unsigned int  fr, unsigned int  ch  )

inline

Constant access to an element of a waveform structure.

Parameters

s	Waveform structure
fr	Frame number
ch	Channel number

Returns

Reference to element

value() [3/8]

mha_complex_t& value ( mha_spec_t *  s, unsigned int  fr, unsigned int  ch  )

inline

Access to an element of a spectrum.

Parameters

s	Spectrum structure
fr	Bin number
ch	Channel number

Returns

Reference to element

value() [4/8]

const mha_complex_t& value ( const mha_spec_t *  s, unsigned int  fr, unsigned int  ch  )

inline

Constant access to an element of a spectrum.

Parameters

s	Spectrum structure
fr	Bin number
ch	Channel number

Returns

Reference to element

value() [5/8]

mha_real_t& value ( mha_wave_t &  s, unsigned int  fr, unsigned int  ch  )

inline

Access to an element of a waveform structure.

Parameters

s	Waveform structure
fr	Frame number
ch	Channel number

Returns

Reference to element

value() [6/8]

const mha_real_t& value ( const mha_wave_t &  s, unsigned int  fr, unsigned int  ch  )

inline

Constant access to an element of a waveform structure.

Parameters

s	Waveform structure
fr	Frame number
ch	Channel number

Returns

Reference to element

value() [7/8]

mha_complex_t& value ( mha_spec_t &  s, unsigned int  fr, unsigned int  ch  )

inline

Access to an element of a spectrum.

Parameters

s	Spectrum structure
fr	Bin number
ch	Channel number

Returns

Reference to element

value() [8/8]

const mha_complex_t& value ( const mha_spec_t &  s, unsigned int  fr, unsigned int  ch  )

inline

Constant access to an element of a spectrum.

Parameters

s	Spectrum structure
fr	Bin number
ch	Channel number

Returns

Reference to element

std_vector_float()

std::vector<float> std_vector_float

(

const mha_wave_t &

w

)

Converts a mha_wave_t structure into a std::vector<float> (interleaved order).

Warning

This function is not real-time safe. Do not use in signal processing thread.

std_vector_vector_float()

std::vector<std::vector<float> > std_vector_vector_float

(

const mha_wave_t &

w

)

Converts a mha_wave_t structure into a std::vector< std::vector<float> > (outer vector represents channels).

Warning

This function is not real-time safe. Do not use in signal processing thread.

std_vector_vector_complex()

std::vector<std::vector<mha_complex_t> > std_vector_vector_complex

(

const mha_spec_t &

w

)

Converts a mha_spec_t structure into a std::vector< std::vector<mha_complex_t> > (outer vector represents channels).

Warning

This function is not real-time safe. Do not use in signal processing thread.

operator+=() [1/4]

mha_wave_t& operator+=	(	mha_wave_t &	self,
		const mha_real_t &	v
	)

Addition operator.

operator+=() [2/4]

mha_wave_t& operator+=	(	mha_wave_t &	self,
		const mha_wave_t &	v
	)

Addition operator.

operator-=() [1/2]

mha_wave_t& operator-=	(	mha_wave_t &	self,
		const mha_wave_t &	v
	)

Subtraction operator.

operator-=() [2/2]

mha_spec_t& operator-=	(	mha_spec_t &	self,
		const mha_spec_t &	v
	)

Subtraction operator.

operator*=() [1/5]

mha_wave_t& operator*=	(	mha_wave_t &	self,
		const mha_real_t &	v
	)

Element-wise multiplication operator.

operator*=() [2/5]

mha_wave_t& operator*=	(	mha_wave_t &	self,
		const mha_wave_t &	v
	)

Element-wise multiplication operator.

operator*=() [3/5]

mha_spec_t& operator*=	(	mha_spec_t &	self,
		const mha_real_t &	v
	)

Element-wise multiplication operator.

operator*=() [4/5]

mha_spec_t& operator*=	(	mha_spec_t &	self,
		const mha_wave_t &	v
	)

Element-wise multiplication operator.

operator*=() [5/5]

mha_spec_t& operator*=	(	mha_spec_t &	self,
		const mha_spec_t &	v
	)

Element-wise multiplication operator.

operator/=() [1/2]

mha_spec_t& operator/=	(	mha_spec_t &	self,
		const mha_spec_t &	v
	)

Element-wise division operator.

operator/=() [2/2]

mha_wave_t& operator/=	(	mha_wave_t &	self,
		const mha_wave_t &	v
	)

Element-wise division operator.

operator+=() [3/4]

mha_spec_t& operator+=	(	mha_spec_t &	self,
		const mha_spec_t &	v
	)

Addition operator.

operator+=() [4/4]

mha_spec_t& operator+=	(	mha_spec_t &	self,
		const mha_real_t &	v
	)

Addition operator.

operator^=()

mha_wave_t& operator^=	(	mha_wave_t &	self,
		const mha_real_t &	arg
	)

Exponent operator.

Warning

This overwrites the xor operator!

copy_channel() [1/2]

void MHASignal::copy_channel	(	mha_spec_t &	self,
		const mha_spec_t &	src,
		unsigned	sch,
		unsigned	dch
	)

Copy one channel of a source signal.

Parameters

self	Destination.
src	Source
sch	Source channel number
dch	Destination channel number

copy_channel() [2/2]

void MHASignal::copy_channel	(	mha_wave_t &	self,
		const mha_wave_t &	src,
		unsigned	src_channel,
		unsigned	dest_channel
	)

Copy one channel of a source signal.

Parameters

self	Destination.
src	Source
src_channel	Source channel number
dest_channel	Destination channel number

rmslevel() [1/2]

mha_real_t MHASignal::rmslevel	(	const mha_spec_t &	s,
		unsigned int	channel,
		unsigned int	fftlen
	)

Return RMS level of a spectrum channel.

Computes the RMS level of the signal in Pascal in the given channel.

Takes into account the the negative frequency bins that are not stored (Central Calibration).

Parameters

s	Input spectrum
channel	Channel number to be tested
fftlen	FFT length (to correctly count the level of the Nyquist bin)

Returns

RMS level in Pa

colored_intensity()

mha_real_t MHASignal::colored_intensity	(	const mha_spec_t &	s,
		unsigned int	channel,
		unsigned int	fftlen,
		mha_real_t *	sqfreq_response = nullptr
	)

Colored spectrum intensity.

computes the squared sum of the spectrum after filtering with the frequency response. Takes into account the negative frequency bins that are not stored (Central Calibration).

Parameters

s	Input spectrum
channel	Channel number to be tested
fftlen	FFT length (to correctly count the level of the Nyquist bin)
sqfreq_response	An array with one squared weighting factor for every fft bin. Array length must be equal to s->num_frames. nullptr can be given for equal weighting of all frequencies.

Returns

sum of squares. Root of this is the colored level in Pa

maxabs() [1/3]

mha_real_t MHASignal::maxabs	(	const mha_spec_t &	s,
		unsigned int	channel
	)

Find maximal absolute value.

Parameters

s	Input signal
channel	Channel to be tested

Returns

maximum absolute value

rmslevel() [2/2]

mha_real_t MHASignal::rmslevel	(	const mha_wave_t &	s,
		unsigned int	channel
	)

Return RMS level of a waveform channel.

Parameters

s	Input waveform signal
channel	Channel number to be tested

Returns

RMS level in Pa

maxabs() [2/3]

mha_real_t MHASignal::maxabs	(	const mha_wave_t &	s,
		unsigned int	channel
	)

Find maximal absolute value.

Parameters

s	Input signal
channel	Channel to be tested

Returns

maximum absolute value

maxabs() [3/3]

mha_real_t MHASignal::maxabs

(

const mha_wave_t &

s

)

Find maximal absolute value.

Parameters

s	Input signal

Returns

maximum absolute value

max()

mha_real_t MHASignal::max

(

const mha_wave_t &

s

)

Find maximal value.

Parameters

s	Input signal

Returns

maximum absolute value

min()

mha_real_t MHASignal::min

(

const mha_wave_t &

s

)

Find minimal value.

Parameters

s	Input signal

Returns

maximum absolute value

sumsqr_channel()

mha_real_t MHASignal::sumsqr_channel	(	const mha_wave_t &	s,
		unsigned int	channel
	)

Calculate sum of squared values in one channel.

Parameters

s	Input signal
channel	Channel

Returns

sumsqr_frame()

mha_real_t MHASignal::sumsqr_frame	(	const mha_wave_t &	s,
		unsigned int	frame
	)

Calculate sum over all channels of squared values.

Parameters

s	Input signal
frame	Frame number

Returns

conjugate()

void conjugate ( mha_spec_t &  self )

inline

Replace (!) the value of this mha_spec_t with its conjugate.