Commit bba34f62 authored by Fabian Bauer's avatar Fabian Bauer
Browse files

mofified FIX_1439_SPEEDUP_elliptic_bpf_48k_generic_upsampledsig to not differ from original

parent 1a15ab5b
Loading
Loading
Loading
Loading
Loading
+1 −1
Original line number Diff line number Diff line
@@ -60,7 +60,7 @@
#define TCX_IMDCT_HEADROOM 1
// *** Currently checking be-nes of all-inactive (should not fail!) https://forge.3gpp.org/rep/sa4/audio/ivas-basop/-/pipelines/50048 -- DONE GREEN!
// *** Currently checking be speedups ( only FIX_1439_SPEEDUP_elliptic_bpf_48k_generic_upsampledsig) for be-nes (should not fail!)
// *** Currently checking be speedups ( only modified version of FIX_1439_SPEEDUP_elliptic_bpf_48k_generic_upsampledsig) for be-nes (should not fail!)
//  --> next test: check nonbe speedups
//
//#define FIX_1439_SPEEDUP_elliptic_bpf_48k_generic_STAGE1
+45 −45
Original line number Diff line number Diff line
@@ -6870,37 +6870,37 @@ void elliptic_bpf_48k_generic_fx(
    {
        i = 4;
        L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                             //0 /*Q_input_fx + 13 + 1 - 3*/
        // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );        //1 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );        //1 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );       /*Q_input_fx + 11 + 13  -15 +2*/
        // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );        //2 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );        //2 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 2], full_band_bpf_fx[3][2] ), 2 ) );       /*Q_input_fx + 11 + 13  -15 +2*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );        //3 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 3], full_band_bpf_fx[3][3] ), 2 ) );       /*Q_input_fx + 11 + 13  -15 +2*/
        // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );            //4 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );            //4 /*Q_input_fx + 13 + 1 - 3*/
        L_tmp[i] = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 4], full_band_bpf_fx[3][4] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
        move32();
        i++;

        // L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                             //1 /*Q_input_fx + 13 + 1 - 3*/
        // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );        //2 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( 0, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );      /*Q_input_fx + 11 + 13  -15 +2*/
        L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                             //1 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );        //2 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );       /*Q_input_fx + 11 + 13  -15 +2*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );        //3 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 2], full_band_bpf_fx[3][2] ), 2 ) );       /*Q_input_fx + 11 + 13  -15 +2*/
        // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );        //4 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );        //4 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 3], full_band_bpf_fx[3][3] ), 2 ) );       /*Q_input_fx + 11 + 13  -15 +2*/
        // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );            //5 /*Q_input_fx + 13 + 1 - 3*/
        L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );            //5 /*Q_input_fx + 13 + 1 - 3*/
        L_tmp[i] = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 4], full_band_bpf_fx[3][4] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
        move32();
        i++;

        FOR( ; i < L_FRAME48k / 3; )
        {
            // L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                             //2 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 );                           // 3 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                           //2 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat(L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX);        //3 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );        //4 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );      //4 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 2], full_band_bpf_fx[3][2] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );        //5 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );      //5 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 3], full_band_bpf_fx[3][3] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );          //6 /*Q_input_fx + 13 + 1 - 3*/
            L_tmp[i] = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 4], full_band_bpf_fx[3][4] ), 2 ) );   /*Q_input_fx + 11 + 13  -15 +2*/
@@ -6908,25 +6908,25 @@ void elliptic_bpf_48k_generic_fx(
            i++;

            L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                           //3 /*Q_input_fx + 13 + 1 - 3*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );        //4 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );      //4 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );        //5 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );      //5 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 2], full_band_bpf_fx[3][2] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );      //6 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 3], full_band_bpf_fx[3][3] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );            //7 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );          //7 /*Q_input_fx + 13 + 1 - 3*/
            L_tmp[i] = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 4], full_band_bpf_fx[3][4] ), 2 ) );   /*Q_input_fx + 11 + 13  -15 +2*/
            move32();
            i++;

            // L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                             //4 /*Q_input_fx + 13 + 1 - 3*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );        //5 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( 0, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );      /*Q_input_fx + 11 + 13  -15 +2*/
            L_tmpX = L_shr( L_mult( input_fx[i - 4], full_band_bpf_fx[0][4] ), 3 );                           //4 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 3], full_band_bpf_fx[0][3] ), 3 ), L_tmpX );      //5 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 1], full_band_bpf_fx[3][1] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 2], full_band_bpf_fx[0][2] ), 3 ), L_tmpX );      //6 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 2], full_band_bpf_fx[3][2] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );        //7 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i - 1], full_band_bpf_fx[0][1] ), 3 ), L_tmpX );      //7 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 3], full_band_bpf_fx[3][3] ), 2 ) );     /*Q_input_fx + 11 + 13  -15 +2*/
            // L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );            //8 /*Q_input_fx + 13 + 1 - 3*/
            L_tmpX = L_add_sat( L_shr( L_mult( input_fx[i], full_band_bpf_fx[0][0] ), 3 ), L_tmpX );          //8 /*Q_input_fx + 13 + 1 - 3*/
            L_tmp[i] = L_sub_sat( L_tmpX, L_shl_sat( Mult_32_16( L_tmp[i - 4], full_band_bpf_fx[3][4] ), 2 ) );   /*Q_input_fx + 11 + 13  -15 +2*/
            move32();
            i++;