LCOV - code coverage report
Current view: top level - src/libavcodec - aacpsdsp_template.c (source / functions) Hit Total Coverage
Test: coverage.info Lines: 122 124 98.4 %
Date: 2017-08-17 10:06:07 Functions: 10 10 100.0 %

          Line data    Source code
       1             : /*
       2             :  * Copyright (c) 2010 Alex Converse <alex.converse@gmail.com>
       3             :  *
       4             :  * This file is part of FFmpeg.
       5             :  *
       6             :  * FFmpeg is free software; you can redistribute it and/or
       7             :  * modify it under the terms of the GNU Lesser General Public
       8             :  * License as published by the Free Software Foundation; either
       9             :  * version 2.1 of the License, or (at your option) any later version.
      10             :  *
      11             :  * FFmpeg is distributed in the hope that it will be useful,
      12             :  * but WITHOUT ANY WARRANTY; without even the implied warranty of
      13             :  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
      14             :  * Lesser General Public License for more details.
      15             :  *
      16             :  * You should have received a copy of the GNU Lesser General Public
      17             :  * License along with FFmpeg; if not, write to the Free Software
      18             :  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
      19             :  *
      20             :  * Note: Rounding-to-nearest used unless otherwise stated
      21             :  *
      22             :  */
      23             : #include <stdint.h>
      24             : 
      25             : #include "config.h"
      26             : #include "libavutil/attributes.h"
      27             : #include "aacpsdsp.h"
      28             : 
      29      116779 : static void ps_add_squares_c(INTFLOAT *dst, const INTFLOAT (*src)[2], int n)
      30             : {
      31             :     int i;
      32     3865899 :     for (i = 0; i < n; i++)
      33     3749120 :         dst[i] += AAC_MADD28(src[i][0], src[i][0], src[i][1], src[i][1]);
      34      116779 : }
      35             : 
      36       66259 : static void ps_mul_pair_single_c(INTFLOAT (*dst)[2], INTFLOAT (*src0)[2], INTFLOAT *src1,
      37             :                                  int n)
      38             : {
      39             :     int i;
      40     2198739 :     for (i = 0; i < n; i++) {
      41     2132480 :         dst[i][0] = AAC_MUL16(src0[i][0], src1[i]);
      42     2132480 :         dst[i][1] = AAC_MUL16(src0[i][1], src1[i]);
      43             :     }
      44       66259 : }
      45             : 
      46       64771 : static void ps_hybrid_analysis_c(INTFLOAT (*out)[2], INTFLOAT (*in)[2],
      47             :                                  const INTFLOAT (*filter)[8][2],
      48             :                                  ptrdiff_t stride, int n)
      49             : {
      50             :     int i, j;
      51             : 
      52      556899 :     for (i = 0; i < n; i++) {
      53      492128 :         INT64FLOAT sum_re = (INT64FLOAT)filter[i][6][0] * in[6][0];
      54      492128 :         INT64FLOAT sum_im = (INT64FLOAT)filter[i][6][0] * in[6][1];
      55             : 
      56     3444896 :         for (j = 0; j < 6; j++) {
      57     2952768 :             INTFLOAT in0_re = in[j][0];
      58     2952768 :             INTFLOAT in0_im = in[j][1];
      59     2952768 :             INTFLOAT in1_re = in[12-j][0];
      60     2952768 :             INTFLOAT in1_im = in[12-j][1];
      61     5905536 :             sum_re += (INT64FLOAT)filter[i][j][0] * (in0_re + in1_re) -
      62     2952768 :                       (INT64FLOAT)filter[i][j][1] * (in0_im - in1_im);
      63     5905536 :             sum_im += (INT64FLOAT)filter[i][j][0] * (in0_im + in1_im) +
      64     2952768 :                       (INT64FLOAT)filter[i][j][1] * (in0_re - in1_re);
      65             :         }
      66             : #if USE_FIXED
      67           0 :         out[i * stride][0] = (int)((sum_re + 0x40000000) >> 31);
      68           0 :         out[i * stride][1] = (int)((sum_im + 0x40000000) >> 31);
      69             : #else
      70      492128 :         out[i * stride][0] = sum_re;
      71      492128 :         out[i * stride][1] = sum_im;
      72             : #endif /* USE_FIXED */
      73             :     }
      74       64771 : }
      75             : 
      76        1622 : static void ps_hybrid_analysis_ileave_c(INTFLOAT (*out)[32][2], INTFLOAT L[2][38][64],
      77             :                                       int i, int len)
      78             : {
      79             :     int j;
      80             : 
      81      100354 :     for (; i < 64; i++) {
      82     3258156 :         for (j = 0; j < len; j++) {
      83     3159424 :             out[i][j][0] = L[0][j][i];
      84     3159424 :             out[i][j][1] = L[1][j][i];
      85             :         }
      86             :     }
      87        1622 : }
      88             : 
      89        3238 : static void ps_hybrid_synthesis_deint_c(INTFLOAT out[2][38][64],
      90             :                                       INTFLOAT (*in)[32][2],
      91             :                                       int i, int len)
      92             : {
      93             :     int n;
      94             : 
      95      200342 :     for (; i < 64; i++) {
      96     6504432 :         for (n = 0; n < len; n++) {
      97     6307328 :             out[0][n][i] = in[i][n][0];
      98     6307328 :             out[1][n][i] = in[i][n][1];
      99             :         }
     100             :     }
     101        3238 : }
     102             : 
     103       50520 : static void ps_decorrelate_c(INTFLOAT (*out)[2], INTFLOAT (*delay)[2],
     104             :                              INTFLOAT (*ap_delay)[PS_QMF_TIME_SLOTS + PS_MAX_AP_DELAY][2],
     105             :                              const INTFLOAT phi_fract[2], const INTFLOAT (*Q_fract)[2],
     106             :                              const INTFLOAT *transient_gain,
     107             :                              INTFLOAT g_decay_slope,
     108             :                              int len)
     109             : {
     110             :     static const INTFLOAT a[] = { Q31(0.65143905753106f),
     111             :                                Q31(0.56471812200776f),
     112             :                                Q31(0.48954165955695f) };
     113             :     INTFLOAT ag[PS_AP_LINKS];
     114             :     int m, n;
     115             : 
     116      202080 :     for (m = 0; m < PS_AP_LINKS; m++)
     117      151560 :         ag[m] = AAC_MUL30(a[m], g_decay_slope);
     118             : 
     119     1667160 :     for (n = 0; n < len; n++) {
     120     1616640 :         INTFLOAT in_re = AAC_MSUB30(delay[n][0], phi_fract[0], delay[n][1], phi_fract[1]);
     121     1616640 :         INTFLOAT in_im = AAC_MADD30(delay[n][0], phi_fract[1], delay[n][1], phi_fract[0]);
     122     6466560 :         for (m = 0; m < PS_AP_LINKS; m++) {
     123     4849920 :             INTFLOAT a_re                = AAC_MUL31(ag[m], in_re);
     124     4849920 :             INTFLOAT a_im                = AAC_MUL31(ag[m], in_im);
     125     4849920 :             INTFLOAT link_delay_re       = ap_delay[m][n+2-m][0];
     126     4849920 :             INTFLOAT link_delay_im       = ap_delay[m][n+2-m][1];
     127     4849920 :             INTFLOAT fractional_delay_re = Q_fract[m][0];
     128     4849920 :             INTFLOAT fractional_delay_im = Q_fract[m][1];
     129     4849920 :             INTFLOAT apd_re = in_re;
     130     4849920 :             INTFLOAT apd_im = in_im;
     131     4849920 :             in_re = AAC_MSUB30(link_delay_re, fractional_delay_re,
     132             :                     link_delay_im, fractional_delay_im);
     133     4849920 :             in_re -= a_re;
     134     4849920 :             in_im = AAC_MADD30(link_delay_re, fractional_delay_im,
     135             :                     link_delay_im, fractional_delay_re);
     136     4849920 :             in_im -= a_im;
     137     4849920 :             ap_delay[m][n+5][0] = apd_re + AAC_MUL31(ag[m], in_re);
     138     4849920 :             ap_delay[m][n+5][1] = apd_im + AAC_MUL31(ag[m], in_im);
     139             :         }
     140     1616640 :         out[n][0] = AAC_MUL16(transient_gain[n], in_re);
     141     1616640 :         out[n][1] = AAC_MUL16(transient_gain[n], in_im);
     142             :     }
     143       50520 : }
     144             : 
     145      141042 : static void ps_stereo_interpolate_c(INTFLOAT (*l)[2], INTFLOAT (*r)[2],
     146             :                                     INTFLOAT h[2][4], INTFLOAT h_step[2][4],
     147             :                                     int len)
     148             : {
     149      141042 :     INTFLOAT h0 = h[0][0];
     150      141042 :     INTFLOAT h1 = h[0][1];
     151      141042 :     INTFLOAT h2 = h[0][2];
     152      141042 :     INTFLOAT h3 = h[0][3];
     153      141042 :     INTFLOAT hs0 = h_step[0][0];
     154      141042 :     INTFLOAT hs1 = h_step[0][1];
     155      141042 :     INTFLOAT hs2 = h_step[0][2];
     156      141042 :     INTFLOAT hs3 = h_step[0][3];
     157             :     int n;
     158             : 
     159     3769746 :     for (n = 0; n < len; n++) {
     160             :         //l is s, r is d
     161     3628704 :         INTFLOAT l_re = l[n][0];
     162     3628704 :         INTFLOAT l_im = l[n][1];
     163     3628704 :         INTFLOAT r_re = r[n][0];
     164     3628704 :         INTFLOAT r_im = r[n][1];
     165     3628704 :         h0 += hs0;
     166     3628704 :         h1 += hs1;
     167     3628704 :         h2 += hs2;
     168     3628704 :         h3 += hs3;
     169     3628704 :         l[n][0] = AAC_MADD30(h0, l_re, h2, r_re);
     170     3628704 :         l[n][1] = AAC_MADD30(h0, l_im, h2, r_im);
     171     3628704 :         r[n][0] = AAC_MADD30(h1, l_re, h3, r_re);
     172     3628704 :         r[n][1] = AAC_MADD30(h1, l_im, h3, r_im);
     173             :     }
     174      141042 : }
     175             : 
     176        7387 : static void ps_stereo_interpolate_ipdopd_c(INTFLOAT (*l)[2], INTFLOAT (*r)[2],
     177             :                                            INTFLOAT h[2][4], INTFLOAT h_step[2][4],
     178             :                                            int len)
     179             : {
     180        7387 :     INTFLOAT h00  = h[0][0],      h10  = h[1][0];
     181        7387 :     INTFLOAT h01  = h[0][1],      h11  = h[1][1];
     182        7387 :     INTFLOAT h02  = h[0][2],      h12  = h[1][2];
     183        7387 :     INTFLOAT h03  = h[0][3],      h13  = h[1][3];
     184        7387 :     INTFLOAT hs00 = h_step[0][0], hs10 = h_step[1][0];
     185        7387 :     INTFLOAT hs01 = h_step[0][1], hs11 = h_step[1][1];
     186        7387 :     INTFLOAT hs02 = h_step[0][2], hs12 = h_step[1][2];
     187        7387 :     INTFLOAT hs03 = h_step[0][3], hs13 = h_step[1][3];
     188             :     int n;
     189             : 
     190      140091 :     for (n = 0; n < len; n++) {
     191             :         //l is s, r is d
     192      132704 :         INTFLOAT l_re = l[n][0];
     193      132704 :         INTFLOAT l_im = l[n][1];
     194      132704 :         INTFLOAT r_re = r[n][0];
     195      132704 :         INTFLOAT r_im = r[n][1];
     196      132704 :         h00 += hs00;
     197      132704 :         h01 += hs01;
     198      132704 :         h02 += hs02;
     199      132704 :         h03 += hs03;
     200      132704 :         h10 += hs10;
     201      132704 :         h11 += hs11;
     202      132704 :         h12 += hs12;
     203      132704 :         h13 += hs13;
     204             : 
     205      132704 :         l[n][0] = AAC_MSUB30_V8(h00, l_re, h02, r_re, h10, l_im, h12, r_im);
     206      132704 :         l[n][1] = AAC_MADD30_V8(h00, l_im, h02, r_im, h10, l_re, h12, r_re);
     207      132704 :         r[n][0] = AAC_MSUB30_V8(h01, l_re, h03, r_re, h11, l_im, h13, r_im);
     208      132704 :         r[n][1] = AAC_MADD30_V8(h01, l_im, h03, r_im, h11, l_re, h13, r_re);
     209             :     }
     210        7387 : }
     211             : 
     212         348 : av_cold void AAC_RENAME(ff_psdsp_init)(PSDSPContext *s)
     213             : {
     214         348 :     s->add_squares            = ps_add_squares_c;
     215         348 :     s->mul_pair_single        = ps_mul_pair_single_c;
     216         348 :     s->hybrid_analysis        = ps_hybrid_analysis_c;
     217         348 :     s->hybrid_analysis_ileave = ps_hybrid_analysis_ileave_c;
     218         348 :     s->hybrid_synthesis_deint = ps_hybrid_synthesis_deint_c;
     219         348 :     s->decorrelate            = ps_decorrelate_c;
     220         348 :     s->stereo_interpolate[0]  = ps_stereo_interpolate_c;
     221         348 :     s->stereo_interpolate[1]  = ps_stereo_interpolate_ipdopd_c;
     222             : 
     223             : #if !USE_FIXED
     224             :     if (ARCH_ARM)
     225             :         ff_psdsp_init_arm(s);
     226             :     if (ARCH_AARCH64)
     227             :         ff_psdsp_init_aarch64(s);
     228             :     if (ARCH_MIPS)
     229             :         ff_psdsp_init_mips(s);
     230             :     if (ARCH_X86)
     231         319 :         ff_psdsp_init_x86(s);
     232             : #endif /* !USE_FIXED */
     233         348 : }

Generated by: LCOV version 1.13