view libtwamr/g_adapt.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 2a265be82195
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             : g_adapt.c
*      Purpose          : gain adaptation for MR795 gain quantization
*
********************************************************************************
*/

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

/*
********************************************************************************
*                         INCLUDE FILES
********************************************************************************
*/
#include "typedef.h"
#include "basic_op.h"
#include "oper_32b.h"
#include "no_count.h"
#include "cnst.h"
#include "gmed_n.h"

/*
********************************************************************************
*                         LOCAL VARIABLES AND TABLES
********************************************************************************
*/
#define LTP_GAIN_THR1 2721 /* 2721 Q13 = 0.3322 ~= 1.0 / (10*log10(2)) */
#define LTP_GAIN_THR2 5443 /* 5443 Q13 = 0.6644 ~= 2.0 / (10*log10(2)) */

/*
********************************************************************************
*                         PUBLIC PROGRAM CODE
********************************************************************************
*/

/*************************************************************************
*
*  Function:   gain_adapt_reset
*  Purpose:    Initializes state memory to zero
*
**************************************************************************
*/
void gain_adapt_reset (GainAdaptState *st)
{
    Word16 i;

    st->onset = 0;
    st->prev_alpha = 0;
    st->prev_gc = 0;

    for (i = 0; i < LTPG_MEM_SIZE; i++)
    {
        st->ltpg_mem[i] = 0;
    }
}

/*************************************************************************
 *
 *  Function:   gain_adapt()
 *  Purpose:    calculate pitch/codebook gain adaptation factor alpha
 *              (and update the adaptor state)
 *
 **************************************************************************
 */
void gain_adapt(
    GainAdaptState *st,  /* i  : state struct                  */
    Word16 ltpg,         /* i  : ltp coding gain (log2()), Q13 */
    Word16 gain_cod,     /* i  : code gain,                Q1  */
    Word16 *alpha        /* o  : gain adaptation factor,   Q15 */
)
{
    Word16 adapt;      /* adaptdation status; 0, 1, or 2       */
    Word16 result;     /* alpha factor, Q13                    */
    Word16 filt;       /* median-filtered LTP coding gain, Q13 */
    Word16 tmp, i;
    
    /* basic adaptation */
    test ();
    if (sub (ltpg, LTP_GAIN_THR1) <= 0)
    {
        adapt = 0;                            move16 ();
    }
    else
    {
        test ();
        if (sub (ltpg, LTP_GAIN_THR2) <= 0)
        {
            adapt = 1;                        move16 ();
        }
        else
        {
            adapt = 2;                        move16 ();
        }
    }

    /*
     * // onset indicator
     * if ((cbGain > onFact * cbGainMem[0]) && (cbGain > 100.0))
     *     onset = 8;
     * else
     *     if (onset)
     *         onset--;
     */
    /* tmp = cbGain / onFact; onFact = 2.0; 200 Q1 = 100.0 */
    tmp = shr_r (gain_cod, 1);
    test (); test ();
    if ((sub (tmp, st->prev_gc) > 0) && sub(gain_cod, 200) > 0)
    {
        st->onset = 8;                            move16 ();
    }
    else
    {
        test ();
        if (st->onset != 0)
        {
            st->onset = sub (st->onset, 1);       move16 ();
        }
    }

    /*
     *  // if onset, increase adaptor state
     *  if (onset && (gainAdapt < 2)) gainAdapt++;
     */
    test(); test ();
    if ((st->onset != 0) && (sub (adapt, 2) < 0))
    {
        adapt = add (adapt, 1);
    }

    st->ltpg_mem[0] = ltpg;                       move16 ();
    filt = gmed_n (st->ltpg_mem, 5);   move16 (); /* function result */

    test ();
    if (adapt == 0)
    {
        test ();
        if (sub (filt, 5443) > 0) /* 5443 Q13 = 0.66443... */
        {
            result = 0; move16 ();
        }
        else
        {
            test ();
            if (filt < 0)
            {
                result = 16384; move16 ();  /* 16384 Q15 = 0.5 */
            }
            else
            {   /* result       =   0.5 - 0.75257499*filt     */
                /* result (Q15) = 16384 - 24660 * (filt << 2) */
                filt = shl (filt, 2); /* Q15 */
                result = sub (16384, mult (24660, filt));
            }
        }
    }
    else
    {
        result = 0; move16 ();
    }
    /*
     *  if (prevAlpha == 0.0) result = 0.5 * (result + prevAlpha);
     */
    test ();
    if (st->prev_alpha == 0)
    {
        result = shr (result, 1);
    }

    /* store the result */
    *alpha = result;                           move16 ();
    
    /* update adapter state memory */
    st->prev_alpha = result;                   move16 ();
    st->prev_gc = gain_cod;                    move16 ();

    for (i = LTPG_MEM_SIZE-1; i > 0; i--)
    {
        st->ltpg_mem[i] = st->ltpg_mem[i-1];   move16 ();
    }
    /* mem[0] is just present for convenience in calling the gmed_n[5]
     * function above. The memory depth is really LTPG_MEM_SIZE-1.
     */
}