Raised Cosine (RC) Pulse Shaper. More...

#include <itpp/comm/pulse_shape.h>

Inheritance diagram for itpp::Raised_Cosine< T1 >:

Public Member Functions
	Raised_Cosine ()
	Constructor.

	Raised_Cosine (double roll_off, int filter_length=6, int upsampling_factor=8)
	Constructor.

virtual	~Raised_Cosine ()
	Destructor.

void	set_pulse_shape (double roll_off_factor, int filter_length=6, int upsampling_factor=8)
	Set pulse shape (roll_off_factor between 0 and 1, filter_length even)

double	get_roll_off (void) const
	Get the roll-off factor.

void	set_pulse_shape (const Vec< double > &impulse_response, int upsampling_factor)
	Set the general impulse response of the FIR filter.

Vec< double >	get_pulse_shape (void) const
	Get the pulse shape.

int	get_upsampling_factor () const
	Get the over sampling factor.

int	get_pulse_length () const
	Get the length of the pulse in number of symbols.

int	get_filter_length () const
	Get the length of the internal FIR filter.

void	shape_symbols (const Vec< T1 > &input, Vec< T1 > &output)
	Shape the input symbols performing upsampling.

Vec< T1 >	shape_symbols (const Vec< T1 > &input)
	Shape the input symbols performing upsampling.

void	shape_samples (const Vec< T1 > &input, Vec< T1 > &output)
	Shape the input samples already upsampled.

Vec< T1 >	shape_samples (const Vec< T1 > &input)
	Shape the input symbols already upsampled.

void	clear (void)
	Clear internal states.

Protected Attributes
double	roll_off_factor
	The roll off factor (i.e. alpha)

Vec< double >	impulse_response
	The impulse resounse of the pulse shaping filter.

MA_Filter< T1, double, T1 >	shaping_filter
	The pulse shaping filter.

int	pulse_length
	Length in symbols.

int	upsampling_factor
	Samples per input symbol.

bool	setup_done
	Ensures that setup is called before any other member function.

Detailed Description

template<class T1>
class itpp::Raised_Cosine< T1 >

Raised Cosine (RC) Pulse Shaper.

Upsamples and shapes symbols as raised cosine pulses with a given roll-off factor $\alpha$ . The raised cosine pulse shape is defined as:

$p(t) = \frac{\sin(\pi t / T)}{\pi t / T} \frac{\cos(\alpha \pi t / T)}{1 - (2 \alpha t / T)^2}$

For more details see e.g. Lee & Messerschmitt, p. 190. Observe that the shaping is done with a FIR filter where the size is given by filter_length * over_sample_factor + 1. The first samples in the output will therefore be zero or small before the memory of the filter is filled.

What is important, when using RC shaping in a transmission system with the AWGN channel, the mean power of the output samples is not normalised, so the channel noise variance (or shaped signal) should be scaled appropriately.

The class is templated as follows: T1 is the type of the input and the output samples. An example of usage is:

#include "itpp/itcomm.h"
Raised_Cosine<double> rc(0.5, 6, 8);
BPSK bpsk;
vec symbols, samples;
symbols = bpsk.modulate_bits(randb(20));
samples = rc.shape_symbols(symbols);

Definition at line 162 of file pulse_shape.h.

Member Function Documentation

void itpp::Pulse_Shape< T1, double , T1 >::set_pulse_shape	(	const Vec< double > &	impulse_response,
		int	upsampling_factor
	)

inherited

Set the general impulse response of the FIR filter.

Observe that the pulse shape must have a duration of an integer number of symbols. Thus the length of the impulse response-1 modulo over sampling is an integer.

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

itpp/comm/pulse_shape.h

Public Member Functions

Protected Attributes

Detailed Description

template<class T1> class itpp::Raised_Cosine< T1 >

Member Function Documentation

template<class T1>
class itpp::Raised_Cosine< T1 >