xref: /frameworks/av/media/libstagefright/codecs/amrnb/enc/src/s10_8pf.cpp

/* ------------------------------------------------------------------
 * Copyright (C) 1998-2009 PacketVideo
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *      http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either
 * express or implied.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 * -------------------------------------------------------------------
 */
/****************************************************************************************
Portions of this file are derived from the following 3GPP standard:

    3GPP TS 26.073
    ANSI-C code for the Adaptive Multi-Rate (AMR) speech codec
    Available from http://www.3gpp.org

(C) 2004, 3GPP Organizational Partners (ARIB, ATIS, CCSA, ETSI, TTA, TTC)
Permission to distribute, modify and use this file under the standard license
terms listed above has been obtained from the copyright holder.
****************************************************************************************/
/*
------------------------------------------------------------------------------



 Pathname: ./audio/gsm-amr/c/src/s10_8pf.c
 Funtions: search_10and8i40

     Date: 04/18/2000

------------------------------------------------------------------------------
 REVISION HISTORY

 Description: Adding pOverflow to the functions to remove global variables.
              These changes are needed for the EPOC releases. Cleaned up code.
              Updated template.

 Description: Changed temp to temp32. When temp was only 16 bits it was not
              holding the 32 bit value returned from the functions. Some
              variables were also being declared as Word16 rather than Word32
              as they were suposed to be.

 Description: Changed copyright year. Removed all calls to math functions by
              inlining them, and removed all unnecessary files in the Include
              section.

 Description: Made the following changes per comments from Phase 2/3 review:
              1. Removed all #defines.
              2. Used a pointer to &codvec[0] instead of array indexing.
              3. Removed multiple data casting in the code.

 Description:
              1. Eliminated unused include files.
              2. Replaced array addressing by pointers, this by taking
                 advantage of the fact that the autocrrelation  matrix is
                 a toeplitz matrix, so r[i][j] = r[j][i], then a single
                 pointer can be used to address a matrix. The use of this
                 is not uniform along the function (due to compiler limitations:
                 handling so many variables in this file) so the use
                 of this is pointer optimizations is limited to places
                 where the ARM compiler provides the lesses numer of cycles
              3. Eliminated use of intermediate variables to accelerate
                 comparisons (like in the nested loops)
              4. Introduced array temp1[], to pre-calculate the elements
                 used in the nested loops, in this way the calculation is
                 not repeated in every loop iteration. This is done for
                 loops i3-i5-i7 and i9
              5. Use array Index[] to store indexes i1:i9, and then use memcpy
                 to update indexes.
              6. Eliminated shifts by modifying the way number are rounded,
                 this does not have any effect in ARM processors but may help
                 other compilers

 Description:
              1. When storing indexes, added memcpy() to support the rates
                 that use this function: 12.2 (already done) and 10.2 (missing).

 Description:  Replaced OSCL mem type functions and eliminated include
               files that now are chosen by OSCL definitions

 Description: Changed round function name to pv_round to avoid conflict with
              round function in C standard library.

 Description:

------------------------------------------------------------------------------
*/

/*----------------------------------------------------------------------------
; INCLUDES
----------------------------------------------------------------------------*/
#include <string.h>

#include "s10_8pf.h"
#include "cnst.h"

/*----------------------------------------------------------------------------
; MACROS
; Define module specific macros here
----------------------------------------------------------------------------*/

/*----------------------------------------------------------------------------
; DEFINES
; Include all pre-processor statements here. Include conditional
; compile variables also.
----------------------------------------------------------------------------*/

/*----------------------------------------------------------------------------
; LOCAL FUNCTION DEFINITIONS
; Function Prototype declaration
----------------------------------------------------------------------------*/

/*----------------------------------------------------------------------------
; LOCAL VARIABLE DEFINITIONS
; Variable declaration - defined here and used outside this module
----------------------------------------------------------------------------*/

/*
------------------------------------------------------------------------------
 FUNCTION NAME: search_10and8i40
------------------------------------------------------------------------------
 INPUT AND OUTPUT DEFINITIONS

 Inputs:
    nbPulse = nbPulses to find (Word16)
    step = step size (Word16)
    nbTracks = nbTracks (Word16)
    dn[] = correlation between target and h[] (Word16)
    rr[][] = matrix of autocorrelation (Word16)
    ipos[] = starting position of each pulse (Word16)
    pos_max[] = Position of maximum dn[] (Word16)
    codvec[] = Algebraic codebook vector (Word16)
    pOverflow = pointer to Overflow flag (Flag)

 Outputs:
    codvec[] = Algebraic codebook vector (Word16)
    pOverflow -> 1 if processing this funvction results in satuaration

 Returns:
    None

 Global Variables Used:
    None

 Local Variables Needed:
    None

------------------------------------------------------------------------------
 FUNCTION DESCRIPTION

 This function searches for the best codevector; It determines the positions
 of the 10/8 pulses in the 40-sample frame.

    search_10and8i40 (10,5,5,dn, rr, ipos, pos_max, codvec);   for GSMEFR
    search_10and8i40 (8, 4,4,dn, rr, ipos, pos_max, codvec);   for 10.2


------------------------------------------------------------------------------
 REQUIREMENTS

 None

------------------------------------------------------------------------------
 REFERENCES

 s10_8pf.c, UMTS GSM AMR speech codec, R99 - Version 3.2.0, March 2, 2001

------------------------------------------------------------------------------
 PSEUDO-CODE

void search_10and8i40 (
    Word16 nbPulse,      // i : nbpulses to find
    Word16 step,         // i :  stepsize
    Word16 nbTracks,     // i :  nbTracks
    Word16 dn[],         // i : correlation between target and h[]
    Word16 rr[][L_CODE], // i : matrix of autocorrelation
    Word16 ipos[],       // i : starting position for each pulse
    Word16 pos_max[],    // i : position of maximum of dn[]
    Word16 codvec[]      // o : algebraic codebook vector
)
{
   Word16 i0, i1, i2, i3, i4, i5, i6, i7, i8, i9;
   Word16 i, j, k, pos, ia, ib;
   Word16 psk, ps, ps0, ps1, ps2, sq, sq2;
   Word16 alpk, alp, alp_16;
   Word16 rrv[L_CODE];
   Word32 s, alp0, alp1, alp2;
   Word16 gsmefrFlag;


   if (sub(nbPulse, 10) == 0)
   {
      gsmefrFlag=1;
   }
   else
   {
      gsmefrFlag=0;
   }

   // fix i0 on maximum of correlation position
   i0 = pos_max[ipos[0]];

   //
   // i1 loop:                                                         *
   //

   // Default value

   psk = -1;
   alpk = 1;
   for (i = 0; i < nbPulse; i++)
   {
      codvec[i] = i;
   }

   for (i = 1; i < nbTracks; i++)
   {
      i1 = pos_max[ipos[1]];
      ps0 = add (dn[i0], dn[i1]);
      alp0 = L_mult (rr[i0][i0], _1_16);
      alp0 = L_mac (alp0, rr[i1][i1], _1_16);
      alp0 = L_mac (alp0, rr[i0][i1], _1_8);

      //
      // i2 and i3 loop
      //

      for (i3 = ipos[3]; i3 < L_CODE; i3 += step)
      {
         s = L_mult (rr[i3][i3], _1_8);       // index incr= step+L_CODE
         s = L_mac (s, rr[i0][i3], _1_4);     // index increment = step
         s = L_mac (s, rr[i1][i3], _1_4);     // index increment = step
         rrv[i3] = pv_round (s);
      }

      // Default value
      sq = -1;
      alp = 1;
      ps = 0;
      ia = ipos[2];
      ib = ipos[3];

      for (i2 = ipos[2]; i2 < L_CODE; i2 += step)
      {
         // index increment = step
         ps1 = add (ps0, dn[i2]);

         // index incr= step+L_CODE
         alp1 = L_mac (alp0, rr[i2][i2], _1_16);

         // index increment = step
         alp1 = L_mac (alp1, rr[i0][i2], _1_8);

         // index increment = step
         alp1 = L_mac (alp1, rr[i1][i2], _1_8);

         for (i3 = ipos[3]; i3 < L_CODE; i3 += step)
         {
            // index increment = step
            ps2 = add (ps1, dn[i3]);

            // index increment = step
            alp2 = L_mac (alp1, rrv[i3], _1_2);

            // index increment = step
            alp2 = L_mac (alp2, rr[i2][i3], _1_8);

            sq2 = mult (ps2, ps2);

            alp_16 = pv_round (alp2);

            s = L_msu (L_mult (alp, sq2), sq, alp_16);

            if (s > 0)
            {
               sq = sq2;
               ps = ps2;
               alp = alp_16;
               ia = i2;
               ib = i3;
            }
         }
      }
      i2 = ia;
      i3 = ib;

        //
        // i4 and i5 loop:
        //

        ps0 = ps;
        alp0 = L_mult (alp, _1_2);

        for (i5 = ipos[5]; i5 < L_CODE; i5 += step)
        {
            s = L_mult (rr[i5][i5], _1_8);
            s = L_mac (s, rr[i0][i5], _1_4);
            s = L_mac (s, rr[i1][i5], _1_4);
            s = L_mac (s, rr[i2][i5], _1_4);
            s = L_mac (s, rr[i3][i5], _1_4);
            rrv[i5] = pv_round (s);
        }

        // Default value
        sq = -1;
        alp = 1;
        ps = 0;
        ia = ipos[4];
        ib = ipos[5];

        for (i4 = ipos[4]; i4 < L_CODE; i4 += step)
        {
            ps1 = add (ps0, dn[i4]);

            alp1 = L_mac (alp0, rr[i4][i4], _1_32);
            alp1 = L_mac (alp1, rr[i0][i4], _1_16);
            alp1 = L_mac (alp1, rr[i1][i4], _1_16);
            alp1 = L_mac (alp1, rr[i2][i4], _1_16);
            alp1 = L_mac (alp1, rr[i3][i4], _1_16);

            for (i5 = ipos[5]; i5 < L_CODE; i5 += step)
            {
                ps2 = add (ps1, dn[i5]);

                alp2 = L_mac (alp1, rrv[i5], _1_4);
                alp2 = L_mac (alp2, rr[i4][i5], _1_16);

                sq2 = mult (ps2, ps2);

                alp_16 = pv_round (alp2);

                s = L_msu (L_mult (alp, sq2), sq, alp_16);

                if (s > 0)
                {
                    sq = sq2;
                    ps = ps2;
                    alp = alp_16;
                    ia = i4;
                    ib = i5;
                }
            }
        }
        i4 = ia;
        i5 = ib;

        //
        // i6 and i7 loop:
        //

        ps0 = ps;
        alp0 = L_mult (alp, _1_2);

        for (i7 = ipos[7]; i7 < L_CODE; i7 += step)
        {
            s = L_mult (rr[i7][i7], _1_16);
            s = L_mac (s, rr[i0][i7], _1_8);
            s = L_mac (s, rr[i1][i7], _1_8);
            s = L_mac (s, rr[i2][i7], _1_8);
            s = L_mac (s, rr[i3][i7], _1_8);
            s = L_mac (s, rr[i4][i7], _1_8);
            s = L_mac (s, rr[i5][i7], _1_8);
            rrv[i7] = pv_round (s);
        }

        // Default value
        sq = -1;
        alp = 1;
        ps = 0;
        ia = ipos[6];
        ib = ipos[7];

        for (i6 = ipos[6]; i6 < L_CODE; i6 += step)
        {
            ps1 = add (ps0, dn[i6]);

            alp1 = L_mac (alp0, rr[i6][i6], _1_64);
            alp1 = L_mac (alp1, rr[i0][i6], _1_32);
            alp1 = L_mac (alp1, rr[i1][i6], _1_32);
            alp1 = L_mac (alp1, rr[i2][i6], _1_32);
            alp1 = L_mac (alp1, rr[i3][i6], _1_32);
            alp1 = L_mac (alp1, rr[i4][i6], _1_32);
            alp1 = L_mac (alp1, rr[i5][i6], _1_32);

            for (i7 = ipos[7]; i7 < L_CODE; i7 += step)
            {
                ps2 = add (ps1, dn[i7]);

                alp2 = L_mac (alp1, rrv[i7], _1_4);
                alp2 = L_mac (alp2, rr[i6][i7], _1_32);

                sq2 = mult (ps2, ps2);

                alp_16 = pv_round (alp2);

                s = L_msu (L_mult (alp, sq2), sq, alp_16);

                if (s > 0)
                {
                    sq = sq2;
                    ps = ps2;
                    alp = alp_16;
                    ia = i6;
                    ib = i7;
                }
            }
        }
        i6 = ia;
        i7 = ib;

        // now finished searching a set of 8 pulses

        if(gsmefrFlag != 0){
           // go on with the two last pulses for GSMEFR
           //
           // i8 and i9 loop:
           //

           ps0 = ps;
           alp0 = L_mult (alp, _1_2);

           for (i9 = ipos[9]; i9 < L_CODE; i9 += step)
           {
              s = L_mult (rr[i9][i9], _1_16);
              s = L_mac (s, rr[i0][i9], _1_8);
              s = L_mac (s, rr[i1][i9], _1_8);
              s = L_mac (s, rr[i2][i9], _1_8);
              s = L_mac (s, rr[i3][i9], _1_8);
              s = L_mac (s, rr[i4][i9], _1_8);
              s = L_mac (s, rr[i5][i9], _1_8);
              s = L_mac (s, rr[i6][i9], _1_8);
              s = L_mac (s, rr[i7][i9], _1_8);
              rrv[i9] = pv_round (s);
           }

           // Default value
           sq = -1;
           alp = 1;
           ps = 0;
           ia = ipos[8];
           ib = ipos[9];

           for (i8 = ipos[8]; i8 < L_CODE; i8 += step)
           {
              ps1 = add (ps0, dn[i8]);

              alp1 = L_mac (alp0, rr[i8][i8], _1_128);
              alp1 = L_mac (alp1, rr[i0][i8], _1_64);
              alp1 = L_mac (alp1, rr[i1][i8], _1_64);
              alp1 = L_mac (alp1, rr[i2][i8], _1_64);
              alp1 = L_mac (alp1, rr[i3][i8], _1_64);
              alp1 = L_mac (alp1, rr[i4][i8], _1_64);
              alp1 = L_mac (alp1, rr[i5][i8], _1_64);
              alp1 = L_mac (alp1, rr[i6][i8], _1_64);
              alp1 = L_mac (alp1, rr[i7][i8], _1_64);

              for (i9 = ipos[9]; i9 < L_CODE; i9 += step)
              {
                 ps2 = add (ps1, dn[i9]);

                 alp2 = L_mac (alp1, rrv[i9], _1_8);
                 alp2 = L_mac (alp2, rr[i8][i9], _1_64);

                 sq2 = mult (ps2, ps2);

                 alp_16 = pv_round (alp2);

                 s = L_msu (L_mult (alp, sq2), sq, alp_16);

                 if (s > 0)
                 {
                    sq = sq2;
                    ps = ps2;
                    alp = alp_16;
                    ia = i8;
                    ib = i9;
                 }
              }
           }
        } // end  gsmefrFlag

        //
        // test and memorise if this combination is better than the last one/
        //

        s = L_msu (L_mult (alpk, sq), psk, alp);

        if (s > 0)
        {
            psk = sq;
            alpk = alp;
            codvec[0] = i0;
            codvec[1] = i1;
            codvec[2] = i2;
            codvec[3] = i3;
            codvec[4] = i4;
            codvec[5] = i5;
            codvec[6] = i6;
            codvec[7] = i7;

            if (gsmefrFlag != 0)
            {
               codvec[8] = ia;
               codvec[9] = ib;
            }
        }

        //
        // Cyclic permutation of i1,i2,i3,i4,i5,i6,i7,(i8 and i9)/
        //

        pos = ipos[1];
        for (j = 1, k = 2; k < nbPulse; j++, k++)
        {
            ipos[j] = ipos[k];
        }
        ipos[sub(nbPulse,1)] = pos;
   } // end 1..nbTracks  loop
}

------------------------------------------------------------------------------
 RESOURCES USED [optional]

 When the code is written for a specific target processor the
 the resources used should be documented below.

 HEAP MEMORY USED: x bytes

 STACK MEMORY USED: x bytes

 CLOCK CYCLES: (cycle count equation for this function) + (variable
                used to represent cycle count for each subroutine
                called)
     where: (cycle count variable) = cycle count for [subroutine
                                     name]

------------------------------------------------------------------------------
 CAUTION [optional]
 [State any special notes, constraints or cautions for users of this function]

------------------------------------------------------------------------------
*/


/*----------------------------------------------------------------------------
; FUNCTION CODE
----------------------------------------------------------------------------*/
void search_10and8i40(
    Word16 nbPulse,      /* i : nbpulses to find                       */
    Word16 step,         /* i : stepsize                               */
    Word16 nbTracks,     /* i : nbTracks                               */
    Word16 dn[],         /* i : correlation between target and h[]     */
    Word16 rr[][L_CODE], /* i : matrix of autocorrelation              */
    Word16 ipos[],       /* i : starting position for each pulse       */
    Word16 pos_max[],    /* i : position of maximum of dn[]            */
    Word16 codvec[],     /* o : algebraic codebook vector              */
    Flag   *pOverflow    /* i/o : overflow flag                        */
)
{
    Word16 i0, i1, i2, i3, i4, i5, i6, i7, i9;
    Word16 i, j, k/*, m*/;
    Word16 pos, ia, ib;
    Word16 psk;
    Word16 sq, sq2;
    Word16 alpk, alp, alp_16;
    Word32 s;
    Word32 alp0, alp1, alp2;
    Word16 gsmefrFlag;
    Word16 *p_codvec = codvec;
    Word16  *p_temp2;

    Word16  temp1[2*L_CODE];
    Word16  *p_temp1;
    Word16  ps2;
    Word16  ps1;
    Word16  ps;
    Word16 ps0;

    Word16  index[10];

    OSCL_UNUSED_ARG(pOverflow);

    if (nbPulse == 10)
    {
        gsmefrFlag = 1;
    }
    else
    {
        gsmefrFlag = 0;
    }

    /* fix i0 on maximum of correlation position */
    i0 = pos_max[ipos[0]];
    index[0] = i0;
    /*------------------------------------------------------------------*
    * i1 loop:                                                         *
    *------------------------------------------------------------------*/

    /* Default value */
    psk = -1;
    alpk = 1;
    for (i = 0; i < nbPulse; i++)
    {
        *(p_codvec++) = i;
    }

    for (i = 1; i < nbTracks; i++)
    {
        i1 = pos_max[ipos[1]];
        index[1] = i1;

        /* ps0 = add (dn[i0], dn[i1], pOverflow);*/
        ps0 = (Word16)((Word32) dn[i0] + dn[i1]);

        /* alp0 = L_mult (rr[i0][i0], _1_16, pOverflow); */
        alp0 = (Word32) rr[i0][i0] << 12;

        /* alp0 = L_mac (alp0, rr[i1][i1], _1_16, pOverflow); */
        alp0 += (Word32) rr[i1][i1] << 12;

        /* alp0 = L_mac (alp0, rr[i0][i1], _1_8, pOverflow); */
        alp0 += (Word32) rr[i0][i1] << 13;
        alp0 += 0x00008000L;

        /*----------------------------------------------------------------*
        * i2 and i3 loop:                                                *
        *----------------------------------------------------------------*/

        p_temp1 = temp1;
        for (i3 = ipos[3]; i3 < L_CODE; i3 += step)
        {
            p_temp2 = &rr[i3][0];
            s  = (Word32) * (p_temp2 + i3) >> 1;
            s += (Word32) * (p_temp2 + i0);
            s += (Word32) * (p_temp2 + i1);
            *(p_temp1++) = ps0 + dn[i3];
            *(p_temp1++) = (Word16)((s + 2) >> 2);
        }

        /* Default value */
        sq = -1;
        alp = 1;
        ps = 0;
        ia = ipos[2];
        ib = ipos[3];

        s = (alp0 >> 12);

        for (j = ipos[2]; j < L_CODE; j += step)
        {
            /* index increment = step  */
            p_temp2 = &rr[j][0];

            alp1 = (s + (Word32) * (p_temp2 + j)) >> 1;

            alp1 += (Word32) * (p_temp2 + i0);

            alp1 += (Word32) * (p_temp2 + i1);

            p_temp1 = temp1;
            ps1 = dn[j];


            for (i3 = ipos[3]; i3 < L_CODE; i3 += step)
            {
                /* index increment = step */
                ps2 = ps1 + *(p_temp1++);

                sq2 = (Word16)(((Word32) ps2 * ps2) >> 15);

                alp2 = (alp1 + p_temp2[i3]) >> 2;
                alp2 = (alp2 + *(p_temp1++)) >> 1;  /*  alp2 is always > 0  */
                if (((Word32) sq2 * alp) > ((Word32) sq * alp2))
                {
                    sq = sq2;
                    ps = ps2;
                    alp = (Word16)alp2;
                    ia = j;
                    ib = i3;
                }
            }

        }
        i2 = ia;
        i3 = ib;
        index[2] = ia;
        index[3] = ib;

        /*----------------------------------------------------------------*
        * i4 and i5 loop:                                                *
        *----------------------------------------------------------------*/

        alp0 = ((Word32) alp << 15) + 0x00008000L;
        p_temp1 = temp1;

        for (i5 = ipos[5]; i5 < L_CODE; i5 += step)
        {
            p_temp2 = &rr[i5][0];
            s = (Word32) * (p_temp2 + i5) >> 1;
            s += (Word32) * (p_temp2 + i0);
            s += (Word32) * (p_temp2 + i1);
            s += (Word32) * (p_temp2 + i2);
            s += (Word32) * (p_temp2 + i3);

            *(p_temp1++) = ps + dn[i5];
            *(p_temp1++) = (Word16)((s + 2) >> 2);
        }

        /* Default value */
        sq = -1;
        alp = 1;
        ps = 0;
        ia = ipos[4];
        ib = ipos[5];

        for (j = ipos[4]; j < L_CODE; j += step)
        {
            /* ps1 = add (ps0, dn[i4], pOverflow); */
            p_temp2 = &rr[j][0];

            /* alp1 = L_mac (alp0, rr[i4][i4], _1_32, pOverflow); */
            alp1 = alp0 + ((Word32) * (p_temp2 + j) << 11);

            /* alp1 = L_mac (alp1, rr[i0][i4], _1_16, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i0) << 12;

            /* alp1 = L_mac (alp1, rr[i1][i4], _1_16, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i1) << 12;

            /* alp1 = L_mac (alp1, rr[i2][i4], _1_16, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i2) << 12;

            /* alp1 = L_mac (alp1, rr[i3][i4], _1_16, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i3) << 12;

            p_temp1 = temp1;
            ps1 =  dn[j];

            for (i5 = ipos[5]; i5 < L_CODE; i5 += step)
            {
                ps2 = ps1 + *(p_temp1++);

                alp2 = alp1 + ((Word32) * (p_temp2 + i5) << 12);

                alp_16 = (Word16)((alp2 + ((Word32) * (p_temp1++) << 14)) >> 16);
                sq2 = (Word16)(((Word32) ps2 * ps2) >> 15);

                if (((Word32) sq2 * alp) > ((Word32) sq * alp_16))
                {
                    sq = sq2;
                    ps = ps2;
                    alp = alp_16;
                    ia = j;
                    ib = i5;
                }

            }
        }
        i4 = ia;
        i5 = ib;
        index[4] = ia;
        index[5] = ib;

        /*----------------------------------------------------------------*
        * i6 and i7 loop:                                                *
        *----------------------------------------------------------------*/

        alp0 = ((Word32) alp << 15) + 0x00008000L;

        p_temp1 = temp1;

        for (i7 = ipos[7]; i7 < L_CODE; i7 += step)
        {
            s = (Word32) rr[i7][i7] >> 1;
            s += (Word32) rr[i0][i7];
            s += (Word32) rr[i1][i7];
            s += (Word32) rr[i2][i7];
            s += (Word32) rr[i3][i7];
            s += (Word32) rr[i4][i7];
            s += (Word32) rr[i5][i7];
            *(p_temp1++) = ps + dn[i7];
            *(p_temp1++) = (Word16)((s + 4) >> 3);
        }


        /* Default value */
        sq = -1;
        alp = 1;
        ps = 0;
        ia = ipos[6];
        ib = ipos[7];

        for (j = ipos[6]; j < L_CODE; j += step)
        {
            /* ps1 = add (ps0, dn[i6], pOverflow); */

            p_temp2 = (Word16 *) & rr[j];

            /* alp1 = L_mac (alp0, rr[i6][i6], _1_64, pOverflow); */
            alp1 = alp0 + ((Word32) * (p_temp2 + j) << 10);

            /* alp1 = L_mac (alp1, rr[i0][i6], _1_32, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i0) << 11;


            /* alp1 = L_mac (alp1, rr[i1][i6], _1_32, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i1) << 11;

            /* alp1 = L_mac (alp1, rr[i2][i6], _1_32, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i2) << 11;

            /* alp1 = L_mac (alp1, rr[i3][i6], _1_32, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i3) << 11;

            /* alp1 = L_mac (alp1, rr[i4][i6], _1_32, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i4) << 11;

            /* alp1 = L_mac (alp1, rr[i5][i6], _1_32, pOverflow); */
            alp1 += (Word32) * (p_temp2 + i5) << 11;

            p_temp1 = temp1;
            ps1 = dn[j];

            for (i7 = ipos[7]; i7 < L_CODE; i7 += step)
            {
                ps2 = ps1 + *(p_temp1++);

                alp2 = alp1 + ((Word32) * (p_temp2 + i7) << 11);

                alp_16 = (Word16)((alp2 + ((Word32) * (p_temp1++) << 14)) >> 16);

                sq2 = (Word16)(((Word32) ps2 * ps2) >> 15);

                if (((Word32) sq2 * alp) > ((Word32) sq * alp_16))
                {
                    sq = sq2;
                    ps = ps2;
                    alp = alp_16;
                    ia = j;
                    ib = i7;
                }
            }
        }

        i6 = ia;
        i7 = ib;
        index[6] = ia;
        index[7] = ib;

        /* now finished searching a set of 8 pulses */

        if (gsmefrFlag != 0)
        {
            /* go on with the two last pulses for GSMEFR                      */
            /*----------------------------------------------------------------*
            * i8 and i9 loop:                                                *
            *----------------------------------------------------------------*/

            alp0 = ((Word32) alp << 15) + 0x00008000L;

            p_temp1 = temp1;

            for (i9 = ipos[9]; i9 < L_CODE; i9 += step)
            {
                s = (Word32) rr[i9][i9] >> 1;
                s += (Word32) rr[i0][i9];
                s += (Word32) rr[i1][i9];
                s += (Word32) rr[i2][i9];
                s += (Word32) rr[i3][i9];
                s += (Word32) rr[i4][i9];
                s += (Word32) rr[i5][i9];
                s += (Word32) rr[i6][i9];
                s += (Word32) rr[i7][i9];

                *(p_temp1++) = ps + dn[i9];
                *(p_temp1++) = (Word16)((s + 4) >> 3);
            }

            /* Default value */
            sq = -1;
            alp = 1;
            ps = 0;
            ia = ipos[8];
            ib = ipos[9];

            for (j = ipos[8]; j < L_CODE; j += step)
            {
                /* ps1 = add (ps0, dn[i8], pOverflow); */
                p_temp2 = &rr[j][0];

                /* alp1 = L_mac (alp0, rr[i8][i8], _1_128, pOverflow); */
                alp1 = alp0 + ((Word32) * (p_temp2 + j) << 9);

                /* alp1 = L_mac (alp1, rr[i0][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i0][j] << 10;

                /* alp1 = L_mac (alp1, rr[i1][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i1][j] << 10;

                /* alp1 = L_mac (alp1, rr[i2][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i2][j] << 10;

                /* alp1 = L_mac (alp1, rr[i3][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i3][j] << 10;

                /* alp1 = L_mac (alp1, rr[i4][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i4][j] << 10;

                /* alp1 = L_mac (alp1, rr[i5][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i5][j] << 10;

                /* alp1 = L_mac (alp1, rr[i6][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i6][j] << 10;

                /* alp1 = L_mac (alp1, rr[i7][i8], _1_64, pOverflow); */
                alp1 += (Word32) rr[i7][j] << 10;

                p_temp1 = temp1;
                ps1 = dn[j];

                for (i9 = ipos[9]; i9 < L_CODE; i9 += step)
                {
                    /* ps2 = add (ps1, dn[i9], pOverflow); */
                    ps2 = ps1 + *(p_temp1++);

                    /* sq2 = mult (ps2, ps2, pOverflow); */
                    sq2 = (Word16)(((Word32) ps2 * ps2) >> 15);

                    /* alp2 = L_mac (alp1, rrv[i9], _1_8, pOverflow); */
                    alp2 = alp1 + ((Word32) * (p_temp2 + i9) << 10) ;

                    /* alp2 = L_mac (alp2, rr[i8][i9], _1_64, pOverflow); */
                    alp_16 = (Word16)((alp2 + ((Word32) * (p_temp1++) << 13)) >> 16);

                    if (((Word32) sq2 * alp) > ((Word32) sq * alp_16))
                    {
                        sq = sq2;
                        ps = ps2;
                        alp = alp_16;
                        ia = j;
                        ib = i9;
                    }
                }
            }

            index[8] = ia;
            index[9] = ib;

        }/* end  gsmefrFlag */

        /*----------------------------------------------------------------  *
         * test and memorise if this combination is better than the last one.*
         *----------------------------------------------------------------*/

        if (((Word32) alpk * sq) > ((Word32) psk * alp))
        {
            psk = sq;
            alpk = alp;

            if (gsmefrFlag != 0)
            {
                memcpy(codvec, index, (2*NB_TRACK)*sizeof(*index));
            }
            else
            {
                memcpy(codvec, index, (2*NB_TRACK_MR102)*sizeof(*index));
            }

        }
        /*----------------------------------------------------------------*
        * Cyclic permutation of i1,i2,i3,i4,i5,i6,i7,(i8 and i9).          *
        *----------------------------------------------------------------*/

        pos = ipos[1];
        for (j = 1, k = 2; k < nbPulse; j++, k++)
        {
            ipos[j] = ipos[k];
        }
        ipos[nbPulse-1] = pos;
    } /* end 1..nbTracks  loop*/
}