view libtwamr/bgnscd.c @ 581:e2d5cad04cbf

libgsmhr1 RxFE: store CN R0+LPC separately from speech In the original GSM 06.06 code the ECU for speech mode is entirely separate from the CN generator, maintaining separate state. (The main intertie between them is the speech vs CN state variable, distinguishing between speech and CN BFIs, in addition to the CN-specific function of distinguishing between initial and update SIDs.) In the present RxFE implementation I initially thought that we could use the same saved_frame buffer for both ECU and CN, overwriting just the first 4 params (R0 and LPC) when a valid SID comes in. However, I now realize it was a bad idea: the original code has a corner case (long sequence of speech-mode BFIs to put the ECU in state 6, then SID and CN-mode BFIs, then a good speech frame) that would be broken by that buffer reuse approach. We could eliminate this corner case by resetting the ECU state when passing through a CN insertion period, but doing so would needlessly increase the behavioral diffs between GSM 06.06 and our version. Solution: use a separate CN-specific buffer for CN R0+LPC parameters, and match the behavior of GSM 06.06 code in this regard.
author Mychaela Falconia <falcon@freecalypso.org>
date Thu, 13 Feb 2025 10:02:45 +0000
parents 8dfb7cbe6b59
children
line wrap: on
line source

/*************************************************************************
*
*      GSM AMR-NB speech codec   R98   Version 7.6.0   December 12, 2001
*                                R99   Version 3.3.0                
*                                REL-4 Version 4.1.0                
*
********************************************************************************
*
*      File             : bgnscd.c
*      Purpose          : Background noise source charateristic detector (SCD)
*
********************************************************************************
*/

/*
********************************************************************************
*                         MODULE INCLUDE FILE AND VERSION ID
********************************************************************************
*/
#include "namespace.h"
#include "bgnscd.h"

#include "typedef.h"
#include "basic_op.h"
#include "no_count.h"
#include "cnst.h"
#include "memops.h"
#include "gmed_n.h"
#include "sqrt_l.h"

/*
********************************************************************************
*                         LOCAL VARIABLES AND TABLES
********************************************************************************
*/
/*-----------------------------------------------------------------*
 *   Decoder constant parameters (defined in "cnst.h")             *
 *-----------------------------------------------------------------*
 *   L_FRAME       : Frame size.                                   *
 *   L_SUBFR       : Sub-frame size.                               *
 *-----------------------------------------------------------------*/

/*
********************************************************************************
*                         PUBLIC PROGRAM CODE
********************************************************************************
*/
/*
**************************************************************************
*
*  Function    : Bgn_scd_reset
*  Purpose     : Resets state memory
*
**************************************************************************
*/
void Bgn_scd_reset (Bgn_scdState *state)
{
   /* Static vectors to zero */
   Set_zero (state->frameEnergyHist, L_ENERGYHIST);

   /* Initialize hangover handling */
   state->bgHangover = 0;
}

/*
**************************************************************************
*
*  Function    : Bgn_scd
*  Purpose     : Charaterice synthesis speech and detect background noise
*  Returns     : background noise decision; 0 = no bgn, 1 = bgn
*
**************************************************************************
*/
Word16 Bgn_scd (Bgn_scdState *st,      /* i : State variables for bgn SCD */
                Word16 ltpGainHist[],  /* i : LTP gain history            */
                Word16 speech[],       /* o : synthesis speech frame      */
                Word16 *voicedHangover /* o : # of frames after last 
                                              voiced frame                */
                )
{
   Word16 i;
   Word16 prevVoiced, inbgNoise;
   Word16 temp;
   Word16 ltpLimit, frameEnergyMin;
   Word16 currEnergy, noiseFloor, maxEnergy, maxEnergyLastPart;
   Word32 s;
   
   /* Update the inBackgroundNoise flag (valid for use in next frame if BFI) */
   /* it now works as a energy detector floating on top                      */ 
   /* not as good as a VAD.                                                  */

   currEnergy = 0;                                   move16 ();
   s = (Word32) 0;                                   move32 ();

   for (i = 0; i < L_FRAME; i++)
   {
       s = L_mac (s, speech[i], speech[i]);
   }

   s = L_shl(s, 2);  

   currEnergy = extract_h (s);

   frameEnergyMin = 32767;                     move16 ();

   for (i = 0; i < L_ENERGYHIST; i++)
   {
      test ();
      if (sub(st->frameEnergyHist[i], frameEnergyMin) < 0)
         frameEnergyMin = st->frameEnergyHist[i];           move16 ();
   }

   noiseFloor = shl (frameEnergyMin, 4); /* Frame Energy Margin of 16 */

   maxEnergy = st->frameEnergyHist[0];               move16 ();
   for (i = 1; i < L_ENERGYHIST-4; i++)
   {
      test ();
      if ( sub (maxEnergy, st->frameEnergyHist[i]) < 0)
      {
         maxEnergy = st->frameEnergyHist[i];         move16 ();
      }
   }
   
   maxEnergyLastPart = st->frameEnergyHist[2*L_ENERGYHIST/3]; move16 ();
   for (i = 2*L_ENERGYHIST/3+1; i < L_ENERGYHIST; i++)
   {
      test ();
      if ( sub (maxEnergyLastPart, st->frameEnergyHist[i] ) < 0)
      {
         maxEnergyLastPart = st->frameEnergyHist[i]; move16 ();     
      }
   }

   inbgNoise = 0;        /* false */                 move16 (); 

   /* Do not consider silence as noise */
   /* Do not consider continuous high volume as noise */
   /* Or if the current noise level is very low */
   /* Mark as noise if under current noise limit */
   /* OR if the maximum energy is below the upper limit */

   test (); test (); test (); test (); test (); 
   if ( (sub(maxEnergy, LOWERNOISELIMIT) > 0) &&
        (sub(currEnergy, FRAMEENERGYLIMIT) < 0) &&
        (sub(currEnergy, LOWERNOISELIMIT) > 0) &&
        ( (sub(currEnergy, noiseFloor) < 0) ||
          (sub(maxEnergyLastPart, UPPERNOISELIMIT) < 0)))
   {
      test ();
      if (sub(add(st->bgHangover, 1), 30) > 0)
      {
         st->bgHangover = 30;                         move16 ();
      } else
      {
         st->bgHangover = add(st->bgHangover, 1);
      }
   }
   else
   {
      st->bgHangover = 0;                             move16 ();    
   }
   
   /* make final decision about frame state , act somewhat cautiosly */
   test ();
   if (sub(st->bgHangover,1) > 0)
      inbgNoise = 1;       /* true  */               move16 ();  

   for (i = 0; i < L_ENERGYHIST-1; i++)
   {
      st->frameEnergyHist[i] = st->frameEnergyHist[i+1]; move16 ();
   }
   st->frameEnergyHist[L_ENERGYHIST-1] = currEnergy;              move16 ();
   
   /* prepare for voicing decision; tighten the threshold after some 
      time in noise */
   ltpLimit = 13926;             /* 0.85  Q14 */     move16 (); 
   test ();
   if (sub(st->bgHangover, 8) > 0)
   {
      ltpLimit = 15565;          /* 0.95  Q14 */     move16 ();
   }
   test ();
   if (sub(st->bgHangover, 15) > 0)
   {
      ltpLimit = 16383;          /* 1.00  Q14 */     move16 ();
   }

   /* weak sort of voicing indication. */
   prevVoiced = 0;        /* false */                move16 ();
   test ();

   if (sub(gmed_n(&ltpGainHist[4], 5), ltpLimit) > 0)
   {
      prevVoiced = 1;     /* true  */                move16 ();
   }
   test ();   
   if (sub(st->bgHangover, 20) > 0) {
      if (sub(gmed_n(ltpGainHist, 9), ltpLimit) > 0)
      {
         prevVoiced = 1;  /* true  */                move16 ();
      }
      else
      {
         prevVoiced = 0;  /* false  */                move16 ();
      }
   }
   
   test ();
   if (prevVoiced)
   {
      *voicedHangover = 0;                        move16 ();
   }
   else
   {
      temp = add(*voicedHangover, 1);
      test ();
      if (sub(temp, 10) > 0)
      {
         *voicedHangover = 10;                    move16 ();
      }
      else
      {
         *voicedHangover = temp;                  move16 ();
      }
   }

   return inbgNoise;
}