shithub: opus

ref: 8cb54e11ca0663206fec0f9e36c07260f83fdf95
dir: /silk/fixed/silk_find_LTP_FIX.c/

View raw version
/***********************************************************************
Copyright (c) 2006-2011, Skype Limited. All rights reserved. 
Redistribution and use in source and binary forms, with or without 
modification, (subject to the limitations in the disclaimer below) 
are permitted provided that the following conditions are met:
- Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright 
notice, this list of conditions and the following disclaimer in the 
documentation and/or other materials provided with the distribution.
- Neither the name of Skype Limited, nor the names of specific 
contributors, may be used to endorse or promote products derived from 
this software without specific prior written permission.
NO EXPRESS OR IMPLIED LICENSES TO ANY PARTY'S PATENT RIGHTS ARE GRANTED 
BY THIS LICENSE. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND 
CONTRIBUTORS ''AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING,
BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND 
FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 
COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 
INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF 
USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON 
ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
***********************************************************************/

#include "silk_main_FIX.h"
#include "silk_tuning_parameters.h"

/* Head room for correlations                           */
#define LTP_CORRS_HEAD_ROOM                             2

void silk_fit_LTP(
    SKP_int32 LTP_coefs_Q16[ LTP_ORDER ],
    SKP_int16 LTP_coefs_Q14[ LTP_ORDER ]
);

void silk_find_LTP_FIX(
    SKP_int16           b_Q14[ MAX_NB_SUBFR * LTP_ORDER ],              /* O    LTP coefs                                                   */
    SKP_int32           WLTP[ MAX_NB_SUBFR * LTP_ORDER * LTP_ORDER ],   /* O    Weight for LTP quantization                                 */
    SKP_int             *LTPredCodGain_Q7,                              /* O    LTP coding gain                                             */
    const SKP_int16     r_lpc[]  ,                                      /* I    residual signal after LPC signal + state for first 10 ms    */
    const SKP_int       lag[ MAX_NB_SUBFR ],                            /* I    LTP lags                                                    */
    const SKP_int32     Wght_Q15[ MAX_NB_SUBFR ],                       /* I    weights                                                     */
    const SKP_int       subfr_length,                                   /* I    subframe length                                             */
    const SKP_int       nb_subfr,                                       /* I    number of subframes                                         */
    const SKP_int       mem_offset,                                     /* I    number of samples in LTP memory                             */
    SKP_int             corr_rshifts[ MAX_NB_SUBFR ]                    /* O    right shifts applied to correlations                        */
)
{
    SKP_int   i, k, lshift;
    const SKP_int16 *r_ptr, *lag_ptr;
    SKP_int16 *b_Q14_ptr;

    SKP_int32 regu;
    SKP_int32 *WLTP_ptr;
    SKP_int32 b_Q16[ LTP_ORDER ], delta_b_Q14[ LTP_ORDER ], d_Q14[ MAX_NB_SUBFR ], nrg[ MAX_NB_SUBFR ], g_Q26;
    SKP_int32 w[ MAX_NB_SUBFR ], WLTP_max, max_abs_d_Q14, max_w_bits;

    SKP_int32 temp32, denom32;
    SKP_int   extra_shifts;
    SKP_int   rr_shifts, maxRshifts, maxRshifts_wxtra, LZs;
    SKP_int32 LPC_res_nrg, LPC_LTP_res_nrg, div_Q16;
    SKP_int32 Rr[ LTP_ORDER ], rr[ MAX_NB_SUBFR ];
    SKP_int32 wd, m_Q12;
    
    b_Q14_ptr = b_Q14;
    WLTP_ptr  = WLTP;
    r_ptr     = &r_lpc[ mem_offset ];
    for( k = 0; k < nb_subfr; k++ ) {
        lag_ptr = r_ptr - ( lag[ k ] + LTP_ORDER / 2 );

        silk_sum_sqr_shift( &rr[ k ], &rr_shifts, r_ptr, subfr_length ); /* rr[ k ] in Q( -rr_shifts ) */

        /* Assure headroom */
        LZs = silk_CLZ32( rr[k] );
        if( LZs < LTP_CORRS_HEAD_ROOM ) {
            rr[ k ] = SKP_RSHIFT_ROUND( rr[ k ], LTP_CORRS_HEAD_ROOM - LZs );
            rr_shifts += ( LTP_CORRS_HEAD_ROOM - LZs );
        }
        corr_rshifts[ k ] = rr_shifts;
        silk_corrMatrix_FIX( lag_ptr, subfr_length, LTP_ORDER, LTP_CORRS_HEAD_ROOM, WLTP_ptr, &corr_rshifts[ k ] );  /* WLTP_fix_ptr in Q( -corr_rshifts[ k ] ) */

        /* The correlation vector always has lower max abs value than rr and/or RR so head room is assured */
        silk_corrVector_FIX( lag_ptr, r_ptr, subfr_length, LTP_ORDER, Rr, corr_rshifts[ k ] );  /* Rr_fix_ptr   in Q( -corr_rshifts[ k ] ) */
        if( corr_rshifts[ k ] > rr_shifts ) {
            rr[ k ] = SKP_RSHIFT( rr[ k ], corr_rshifts[ k ] - rr_shifts ); /* rr[ k ] in Q( -corr_rshifts[ k ] ) */
        }
        SKP_assert( rr[ k ] >= 0 );

        regu = 1;
        regu = SKP_SMLAWB( regu, rr[ k ], SILK_FIX_CONST( LTP_DAMPING/3, 16 ) );
        regu = SKP_SMLAWB( regu, matrix_ptr( WLTP_ptr, 0, 0, LTP_ORDER ), SILK_FIX_CONST( LTP_DAMPING/3, 16 ) );
        regu = SKP_SMLAWB( regu, matrix_ptr( WLTP_ptr, LTP_ORDER-1, LTP_ORDER-1, LTP_ORDER ), SILK_FIX_CONST( LTP_DAMPING/3, 16 ) );
        silk_regularize_correlations_FIX( WLTP_ptr, &rr[k], regu, LTP_ORDER );

        silk_solve_LDL_FIX( WLTP_ptr, LTP_ORDER, Rr, b_Q16 ); /* WLTP_fix_ptr and Rr_fix_ptr both in Q(-corr_rshifts[k]) */

        /* Limit and store in Q14 */
        silk_fit_LTP( b_Q16, b_Q14_ptr );

        /* Calculate residual energy */
        nrg[ k ] = silk_residual_energy16_covar_FIX( b_Q14_ptr, WLTP_ptr, Rr, rr[ k ], LTP_ORDER, 14 ); /* nrg_fix in Q( -corr_rshifts[ k ] ) */

        /* temp = Wght[ k ] / ( nrg[ k ] * Wght[ k ] + 0.01f * subfr_length ); */
        extra_shifts = SKP_min_int( corr_rshifts[ k ], LTP_CORRS_HEAD_ROOM );
        denom32 = SKP_LSHIFT_SAT32( SKP_SMULWB( nrg[ k ], Wght_Q15[ k ] ), 1 + extra_shifts ) + /* Q( -corr_rshifts[ k ] + extra_shifts ) */
            SKP_RSHIFT( SKP_SMULWB( subfr_length, 655 ), corr_rshifts[ k ] - extra_shifts );    /* Q( -corr_rshifts[ k ] + extra_shifts ) */
        denom32 = SKP_max( denom32, 1 );
        SKP_assert( ((SKP_int64)Wght_Q15[ k ] << 16 ) < SKP_int32_MAX );                        /* Wght always < 0.5 in Q0 */
        temp32 = SKP_DIV32( SKP_LSHIFT( ( SKP_int32 )Wght_Q15[ k ], 16 ), denom32 );            /* Q( 15 + 16 + corr_rshifts[k] - extra_shifts ) */
        temp32 = SKP_RSHIFT( temp32, 31 + corr_rshifts[ k ] - extra_shifts - 26 );              /* Q26 */
        
        /* Limit temp such that the below scaling never wraps around */
        WLTP_max = 0;
        for( i = 0; i < LTP_ORDER * LTP_ORDER; i++ ) {
            WLTP_max = SKP_max( WLTP_ptr[ i ], WLTP_max );
        }
        lshift = silk_CLZ32( WLTP_max ) - 1 - 3; /* keep 3 bits free for vq_nearest_neighbor_fix */
        SKP_assert( 26 - 18 + lshift >= 0 );
        if( 26 - 18 + lshift < 31 ) {
            temp32 = SKP_min_32( temp32, SKP_LSHIFT( ( SKP_int32 )1, 26 - 18 + lshift ) );
        }

        silk_scale_vector32_Q26_lshift_18( WLTP_ptr, temp32, LTP_ORDER * LTP_ORDER ); /* WLTP_ptr in Q( 18 - corr_rshifts[ k ] ) */
        
        w[ k ] = matrix_ptr( WLTP_ptr, LTP_ORDER/2, LTP_ORDER/2, LTP_ORDER ); /* w in Q( 18 - corr_rshifts[ k ] ) */
        SKP_assert( w[k] >= 0 );

        r_ptr     += subfr_length;
        b_Q14_ptr += LTP_ORDER;
        WLTP_ptr  += LTP_ORDER * LTP_ORDER;
    }

    maxRshifts = 0;
    for( k = 0; k < nb_subfr; k++ ) {
        maxRshifts = SKP_max_int( corr_rshifts[ k ], maxRshifts );
    }

    /* Compute LTP coding gain */
    if( LTPredCodGain_Q7 != NULL ) {
        LPC_LTP_res_nrg = 0;
        LPC_res_nrg     = 0;
        SKP_assert( LTP_CORRS_HEAD_ROOM >= 2 ); /* Check that no overflow will happen when adding */
        for( k = 0; k < nb_subfr; k++ ) {
            LPC_res_nrg     = SKP_ADD32( LPC_res_nrg,     SKP_RSHIFT( SKP_ADD32( SKP_SMULWB(  rr[ k ], Wght_Q15[ k ] ), 1 ), 1 + ( maxRshifts - corr_rshifts[ k ] ) ) ); /*  Q( -maxRshifts ) */
            LPC_LTP_res_nrg = SKP_ADD32( LPC_LTP_res_nrg, SKP_RSHIFT( SKP_ADD32( SKP_SMULWB( nrg[ k ], Wght_Q15[ k ] ), 1 ), 1 + ( maxRshifts - corr_rshifts[ k ] ) ) ); /*  Q( -maxRshifts ) */
        }
        LPC_LTP_res_nrg = SKP_max( LPC_LTP_res_nrg, 1 ); /* avoid division by zero */

        div_Q16 = silk_DIV32_varQ( LPC_res_nrg, LPC_LTP_res_nrg, 16 );
        *LTPredCodGain_Q7 = ( SKP_int )SKP_SMULBB( 3, silk_lin2log( div_Q16 ) - ( 16 << 7 ) );

        SKP_assert( *LTPredCodGain_Q7 == ( SKP_int )SKP_SAT16( SKP_MUL( 3, silk_lin2log( div_Q16 ) - ( 16 << 7 ) ) ) );
    }

    /* smoothing */
    /* d = sum( B, 1 ); */
    b_Q14_ptr = b_Q14;
    for( k = 0; k < nb_subfr; k++ ) {
        d_Q14[ k ] = 0;
        for( i = 0; i < LTP_ORDER; i++ ) {
            d_Q14[ k ] += b_Q14_ptr[ i ];
        }
        b_Q14_ptr += LTP_ORDER;
    }

    /* m = ( w * d' ) / ( sum( w ) + 1e-3 ); */
        
    /* Find maximum absolute value of d_Q14 and the bits used by w in Q0 */
    max_abs_d_Q14 = 0;
    max_w_bits    = 0;
    for( k = 0; k < nb_subfr; k++ ) {
        max_abs_d_Q14 = SKP_max_32( max_abs_d_Q14, SKP_abs( d_Q14[ k ] ) );
        /* w[ k ] is in Q( 18 - corr_rshifts[ k ] ) */
        /* Find bits needed in Q( 18 - maxRshifts ) */
        max_w_bits = SKP_max_32( max_w_bits, 32 - silk_CLZ32( w[ k ] ) + corr_rshifts[ k ] - maxRshifts ); 
    }

    /* max_abs_d_Q14 = (5 << 15); worst case, i.e. LTP_ORDER * -SKP_int16_MIN */
    SKP_assert( max_abs_d_Q14 <= ( 5 << 15 ) );

    /* How many bits is needed for w*d' in Q( 18 - maxRshifts ) in the worst case, of all d_Q14's being equal to max_abs_d_Q14 */
    extra_shifts = max_w_bits + 32 - silk_CLZ32( max_abs_d_Q14 ) - 14;
    
    /* Subtract what we got available; bits in output var plus maxRshifts */
    extra_shifts -= ( 32 - 1 - 2 + maxRshifts ); /* Keep sign bit free as well as 2 bits for accumulation */
    extra_shifts = SKP_max_int( extra_shifts, 0 );

    maxRshifts_wxtra = maxRshifts + extra_shifts;
    
    temp32 = SKP_RSHIFT( 262, maxRshifts + extra_shifts ) + 1; /* 1e-3f in Q( 18 - (maxRshifts + extra_shifts) ) */
    wd = 0;
    for( k = 0; k < nb_subfr; k++ ) {
        /* w has at least 2 bits of headroom so no overflow should happen */
        temp32 = SKP_ADD32( temp32,                     SKP_RSHIFT( w[ k ], maxRshifts_wxtra - corr_rshifts[ k ] ) );                    /* Q( 18 - maxRshifts_wxtra ) */
        wd     = SKP_ADD32( wd, SKP_LSHIFT( SKP_SMULWW( SKP_RSHIFT( w[ k ], maxRshifts_wxtra - corr_rshifts[ k ] ), d_Q14[ k ] ), 2 ) ); /* Q( 18 - maxRshifts_wxtra ) */
    }
    m_Q12 = silk_DIV32_varQ( wd, temp32, 12 );

    b_Q14_ptr = b_Q14;
    for( k = 0; k < nb_subfr; k++ ) {
        /* w_fix[ k ] from Q( 18 - corr_rshifts[ k ] ) to Q( 16 ) */
        if( 2 - corr_rshifts[k] > 0 ) {
            temp32 = SKP_RSHIFT( w[ k ], 2 - corr_rshifts[ k ] );
        } else {
            temp32 = SKP_LSHIFT_SAT32( w[ k ], corr_rshifts[ k ] - 2 );
        }

        g_Q26 = SKP_MUL( 
            SKP_DIV32( 
                SILK_FIX_CONST( LTP_SMOOTHING, 26 ), 
                SKP_RSHIFT( SILK_FIX_CONST( LTP_SMOOTHING, 26 ), 10 ) + temp32 ),                                       /* Q10 */ 
            SKP_LSHIFT_SAT32( SKP_SUB_SAT32( ( SKP_int32 )m_Q12, SKP_RSHIFT( d_Q14[ k ], 2 ) ), 4 ) );  /* Q16 */

        temp32 = 0;
        for( i = 0; i < LTP_ORDER; i++ ) {
            delta_b_Q14[ i ] = SKP_max_16( b_Q14_ptr[ i ], 1638 );  /* 1638_Q14 = 0.1_Q0 */
            temp32 += delta_b_Q14[ i ];                          /* Q14 */
        }
        temp32 = SKP_DIV32( g_Q26, temp32 ); /* Q14->Q12 */
        for( i = 0; i < LTP_ORDER; i++ ) {
            b_Q14_ptr[ i ] = SKP_LIMIT_32( ( SKP_int32 )b_Q14_ptr[ i ] + SKP_SMULWB( SKP_LSHIFT_SAT32( temp32, 4 ), delta_b_Q14[ i ] ), -16000, 28000 );
        }
        b_Q14_ptr += LTP_ORDER;
    }
TOC(find_LTP_FIX)
}

void silk_fit_LTP(
    SKP_int32 LTP_coefs_Q16[ LTP_ORDER ],
    SKP_int16 LTP_coefs_Q14[ LTP_ORDER ]
)
{
    SKP_int i;

    for( i = 0; i < LTP_ORDER; i++ ) {
        LTP_coefs_Q14[ i ] = ( SKP_int16 )SKP_SAT16( SKP_RSHIFT_ROUND( LTP_coefs_Q16[ i ], 2 ) );
    }
}