MityDSP::tcSigProcTIDspSupport Class Reference

#include <SigProc/SigTIDspSupport.h>

List of all members.

Static Public Member Functions
static void	gen_w_r2 (float *w, int N)
static void	bit_rev (float *w, int N)
static void	ExtractDualRealFFT (float *FFTData_Combined, float *FFTA, float *FFTB, int N, int NumBins)
static void	compute_bitrev_index (short *index, int nx)
static int	bitrev_index_length (int nx)
static void	ExtractRealFFT (float *FFTData_Combined, float *FFTData, float *Twiddles, int N, int NumBins)
static void	BuildExtractRealFFTTwiddles (float *Twiddles, int N)

---

Detailed Description

The tcSigProcTIDspSupport class provides utility methods useful when using the TI provided DSP libraries.

See also:

tcSigProcTIDspSupport Page

---

Member Function Documentation

void tcSigProcTIDspSupport::gen_w_r2	(	float *	w,
		int	N
	)			[static]

This is a utility method to generate twiddle factors needed for the FFT processing libraries provided by TI.

Parameters:

[out]	w	array for storing twiddle factors
[in]	N	length of the array of data

void tcSigProcTIDspSupport::bit_rev	(	float *	x,
		int	N
	)			[static]

This is a utility method to reverse the order of a set of twiddle factors according to the instructions provided by TI for using the TI DSP libraries.

Parameters:

[in,out]	x	array of data to reorder in bit-reveresed order
[in]	N	length of the array of data

void tcSigProcTIDspSupport::ExtractDualRealFFT	(	float *	FFTData_Combined,
		float *	FFTA,
		float *	FFTB,
		int	N,
		int	NumBins = -1
	)			[static]

This routine will extract FFT spectra from the FFT results of 2 real signals that were combined as I and Q data samples prior to calling a complex FFT function. This utility is useful when only a complex FFT is available for signal processing and real input signals are presented: 2 time sets may be FFT'd using a single complex FFT call.

Parameters:

[in]	FFTData_Combined	I/Q pairs of the FFT data, ordered. Length N complex numbers
[out]	FFTA	complex I/Q results of length NumBins of the spectrum from the real samples
[out]	FFTB	complex I/Q results of length NumBins of the spectrum from the imaginary samples
[in]	N	size of the FFT
[in]	NumBins	Number of I/Q samples to extract. Typically this should be no greater than N/2 as no additional information is useful in the second half of the FFT (it's the complex conjugate of the first half). This allows for faster extraction if only a sub-band is require. If < 0, all N bins will be extracted.

Note:

See http://www.engineeringproductivitytools.com/stuff/T0001/PT10.HTM

void tcSigProcTIDspSupport::compute_bitrev_index	(	short *	index,
		int	nx
	)			[static]

This routine calculates the index for bit reversal of an array of length nx. The length of the index table is 2^(2*ceil(k/2)) where nx = 2^k.

In other words, the length of the index table is:

• for even power of radix: sqrt(nx)
• for odd power of radix: sqrt(2*nx)

Parameters:

[out]	index	indexes for use with DSPF_sp_bitrev_cplx
[in]	nx	length of array to reverse

Note:

this routine was copied from the TI DSP user's guide

int tcSigProcTIDspSupport::bitrev_index_length

(

int

nx

)

[static]

This routine determines the size of the table that will be generated by the compute_bitrev_index.

void tcSigProcTIDspSupport::ExtractRealFFT	(	float *	FFTData_Combined,
		float *	FFTData,
		float *	Twiddles,
		int	N,
		int	NumBins = -1
	)			[static]

This routine will extract FFT spectra from the FFT results of a real signals that was input as I and Q data samples prior to calling a complex FFT function. This utility is useful when only a complex FFT is available for signal processing and a real input signal is presented. The FFT previously run is only N/2 instead of N.

Parameters:

[in]	FFTData_Combined	I/Q pairs of the FFT data, ordered. Length N/2 complex numbers
[out]	FFTData	I/Q results of length NumBins of the resulting spectrum
[out]	Twiddles	array of N complex twiddles = -j*e(-j*2*pi*k/N)
[in]	N	size of the desired FFT (FFT run was N/2)
[in]	NumBins	Number of I/Q samples to extract. Typically this should be no greater than N/2 as no additional information is useful in the second half of the FFT (it's the complex conjugate of the first half). This allows for faster extraction if only a sub-band is require. If < 0, all N bins will be extracted.

Note:

See http://www.engineeringproductivitytools.com/stuff/T0001/PT10.HTM

void tcSigProcTIDspSupport::BuildExtractRealFFTTwiddles	(	float *	Twiddles,
		int	N
	)			[static]

This routine is used to create the twiddles when using ExtractRealFFT

Parameters:

[out]	Twiddles	array of N complex twiddles = -j*e(-j*2*pi*k/N)
[in]	N	size of the desired FFT

Note:

See http://www.engineeringproductivitytools.com/stuff/T0001/PT10.HTM