tcSigProcTIDspSupport

Introduction

The tcSigProcTIDspSupport class provides utility methods that are necessary to leverage the TI Digital Signal Processing libraries for the C6000 series processors. These functions are in many cases copied directly from the TI DSP library user's guide.

See also:

MityDSP::tcSigProcTIDspSupport Class Reference

Example

This is a simple example of tcSigProcTIDspSupport usage. The example shown here computes the FFT spectra from 2 real time domain input signals using the TI provided FFT calls and associate runtime libraries. Additionally, the input data is weighted using a hanning taper. Several static structures are built at initialization, including the windowing taper, a working buffer, twiddle factors for the FFT, and a bit reversal index for sorting the outputs of the FFT (which are in bit-reversed order).

 {

   #define FFT_SIZE 2048 

   float* mpHanningData = new float[FFT_SIZE];       // hanning window
   float* mpTwiddles    = new float[FFT_SIZE*2+8];   // TI FFT twiddle factors
   float* mpWorkBuff    = new float[FFT_SIZE*2+8];   // working buffer
   short* mpInputBuffA  = new float[FFT_SIZE];       // input A (real data, from 16 bit ADC)
   short* mpInputBuffB  = new float[FFT_SIZE];       // input B (real data, from 16 bit ADC)
   float* mpFFTA        = new float[FFT_SIZE*2];     // FFT output for A
   float* mpFFTB        = new float[FFT_SIZE*2];     // FFT output for B
   short* mpBitReverseIndex = new short[tcSigProcTIDspSupport::bitrev_index_length(FFT_SIZE)];

   // build a hanning window
   tcSigProcWinFunc::Hanning(mpHanningData, FFT_SIZE);
   
   // The TI libraries require 8 byte aligned data structures....
   // this can be ignored if using static data structures that
   // are aligned using the #pragma DATA_ALIGN() TI call.
   if ((unsigned int)mpTwiddles & 7)
   {
      mpTwiddles = (float*)(((unsigned int)mpTwiddles+8)&~7);
   }

   // same deal with the working buffer
   if ((unsigned int)mpWorkBuff & 7)
   {
      mpWorkBuff = (float*)(((unsigned int)mpWorkBuff+8)&~7);
   }
   
   tcSigProcTIDspSupport::compute_bitrev_index(mpBitReverseIndex, LOW_RES_FFT_SIZE);

   // Generate the twiddle factors
   tcSigProcTIDspSupport::gen_w_r2(mpTwiddles,FFT_SIZE);
   tcSigProcTIDspSupport::bit_rev(mpTwiddles,FFT_SIZE>>1);

   // ... get the data to be FFT'd

   // move first set into real values of the working buffer
   tcSigProcVector::ToFloat(mpInputBuffA,mpWorkBuff, FFT_SIZE, 1, 2);

   // move second set into complex values of the working buffer
   tcSigProcVector::ToFloat(mpInputBuffB,mpWorkBuff+1, FFT_SIZE, 1, 2);
                               
   // Apply the taper to the real part
      tcSigProcVector::VecAddMul(mpWorkBuff, mpHanningData, 0.0,
                                  mpWorkBuff, FFT_SIZE, 2, 1, 2);
                                  
   // Apply the taper to the imaginary part
   tcSigProcVector::VecAddMul(mpWorkBuff+1, mpHanningData, 0.0,
                              mpWorkBuff+1, FFT_SIZE, 2, 1, 2);

   // do the TI FFT
   DSPF_sp_cfftr2_dit(mpWorkBuff,mpTwiddles,FFT_SIZE);

   // fix the output order to be sorted normally
   DSPF_sp_bitrev_cplx((double*)mpWorkBuff, mpBitReverseIndex, FFT_SIZE);

   // extract the two sets of data
   tcSigProcTIDspSupport::ExtractDualRealFFT(mpWorkBuff, mpFFTA, mpFFTB, 
                                             FFT_SIZE, FFT_SIZE);

   // ... process the results
 }