falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
44748 Commits
32 Branches
293MB
Tree: 2f792cb670
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '2f792cb670'
${ noResults }
FFmpeg/libavfilter/af_volume.c

369 lines
12KB
Raw Blame History 

  1. /*

  2.  * Copyright (c) 2011 Stefano Sabatini

  3.  * Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com>

  4.  *

  5.  * This file is part of Libav.

  6.  *

  7.  * Libav is free software; you can redistribute it and/or

  8.  * modify it under the terms of the GNU Lesser General Public

  9.  * License as published by the Free Software Foundation; either

  10.  * version 2.1 of the License, or (at your option) any later version.

  11.  *

  12.  * Libav is distributed in the hope that it will be useful,

  13.  * but WITHOUT ANY WARRANTY; without even the implied warranty of

  14.  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

  15.  * Lesser General Public License for more details.

  16.  *

  17.  * You should have received a copy of the GNU Lesser General Public

  18.  * License along with Libav; if not, write to the Free Software

  19.  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

  20.  */

  21. 

  22. /**

  23.  * @file

  24.  * audio volume filter

  25.  */

  26. 

  27. #include "libavutil/channel_layout.h"

  28. #include "libavutil/common.h"

  29. #include "libavutil/eval.h"

  30. #include "libavutil/float_dsp.h"

  31. #include "libavutil/intreadwrite.h"

  32. #include "libavutil/opt.h"

  33. #include "libavutil/replaygain.h"

  34. 

  35. #include "audio.h"

  36. #include "avfilter.h"

  37. #include "formats.h"

  38. #include "internal.h"

  39. #include "af_volume.h"

  40. 

  41. static const char * const precision_str[] = {

  42.     "fixed", "float", "double"

  43. };

  44. 

  45. #define OFFSET(x) offsetof(VolumeContext, x)

  46. #define A AV_OPT_FLAG_AUDIO_PARAM

  47. 

  48. static const AVOption options[] = {

  49.     { "volume", "Volume adjustment.",

  50.             OFFSET(volume), AV_OPT_TYPE_DOUBLE, { .dbl = 1.0 }, 0, 0x7fffff, A },

  51.     { "precision", "Mathematical precision.",

  52.             OFFSET(precision), AV_OPT_TYPE_INT, { .i64 = PRECISION_FLOAT }, PRECISION_FIXED, PRECISION_DOUBLE, A, "precision" },

  53.         { "fixed",  "8-bit fixed-point.",     0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_FIXED  }, INT_MIN, INT_MAX, A, "precision" },

  54.         { "float",  "32-bit floating-point.", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_FLOAT  }, INT_MIN, INT_MAX, A, "precision" },

  55.         { "double", "64-bit floating-point.", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_DOUBLE }, INT_MIN, INT_MAX, A, "precision" },

  56.     { "replaygain", "Apply replaygain side data when present",

  57.             OFFSET(replaygain), AV_OPT_TYPE_INT, { .i64 = REPLAYGAIN_DROP }, REPLAYGAIN_DROP, REPLAYGAIN_ALBUM, A, "replaygain" },

  58.         { "drop",   "replaygain side data is dropped", 0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_DROP   }, 0, 0, A, "replaygain" },

  59.         { "ignore", "replaygain side data is ignored", 0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_IGNORE }, 0, 0, A, "replaygain" },

  60.         { "track",  "track gain is preferred",         0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_TRACK  }, 0, 0, A, "replaygain" },

  61.         { "album",  "album gain is preferred",         0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_ALBUM  }, 0, 0, A, "replaygain" },

  62.     { "replaygain_preamp", "Apply replaygain pre-amplification",

  63.             OFFSET(replaygain_preamp), AV_OPT_TYPE_DOUBLE, { .dbl = 0.0 }, -15.0, 15.0, A },

  64.     { "replaygain_noclip", "Apply replaygain clipping prevention",

  65.             OFFSET(replaygain_noclip), AV_OPT_TYPE_INT, { .i64 = 1 }, 0, 1, A },

  66.     { NULL },

  67. };

  68. 

  69. static const AVClass volume_class = {

  70.     .class_name = "volume filter",

  71.     .item_name  = av_default_item_name,

  72.     .option     = options,

  73.     .version    = LIBAVUTIL_VERSION_INT,

  74. };

  75. 

  76. static av_cold int init(AVFilterContext *ctx)

  77. {

  78.     VolumeContext *vol = ctx->priv;

  79. 

  80.     if (vol->precision == PRECISION_FIXED) {

  81.         vol->volume_i = (int)(vol->volume * 256 + 0.5);

  82.         vol->volume   = vol->volume_i / 256.0;

  83.         av_log(ctx, AV_LOG_VERBOSE, "volume:(%d/256)(%f)(%1.2fdB) precision:fixed\n",

  84.                vol->volume_i, vol->volume, 20.0*log(vol->volume)/M_LN10);

  85.     } else {

  86.         av_log(ctx, AV_LOG_VERBOSE, "volume:(%f)(%1.2fdB) precision:%s\n",

  87.                vol->volume, 20.0*log(vol->volume)/M_LN10,

  88.                precision_str[vol->precision]);

  89.     }

  90. 

  91.     return 0;

  92. }

  93. 

  94. static int query_formats(AVFilterContext *ctx)

  95. {

  96.     VolumeContext *vol = ctx->priv;

  97.     AVFilterFormats *formats = NULL;

  98.     AVFilterChannelLayouts *layouts;

  99.     static const enum AVSampleFormat sample_fmts[][7] = {

  100.         /* PRECISION_FIXED */

  101.         {

  102.             AV_SAMPLE_FMT_U8,

  103.             AV_SAMPLE_FMT_U8P,

  104.             AV_SAMPLE_FMT_S16,

  105.             AV_SAMPLE_FMT_S16P,

  106.             AV_SAMPLE_FMT_S32,

  107.             AV_SAMPLE_FMT_S32P,

  108.             AV_SAMPLE_FMT_NONE

  109.         },

  110.         /* PRECISION_FLOAT */

  111.         {

  112.             AV_SAMPLE_FMT_FLT,

  113.             AV_SAMPLE_FMT_FLTP,

  114.             AV_SAMPLE_FMT_NONE

  115.         },

  116.         /* PRECISION_DOUBLE */

  117.         {

  118.             AV_SAMPLE_FMT_DBL,

  119.             AV_SAMPLE_FMT_DBLP,

  120.             AV_SAMPLE_FMT_NONE

  121.         }

  122.     };

  123. 

  124.     layouts = ff_all_channel_layouts();

  125.     if (!layouts)

  126.         return AVERROR(ENOMEM);

  127.     ff_set_common_channel_layouts(ctx, layouts);

  128. 

  129.     formats = ff_make_format_list(sample_fmts[vol->precision]);

  130.     if (!formats)

  131.         return AVERROR(ENOMEM);

  132.     ff_set_common_formats(ctx, formats);

  133. 

  134.     formats = ff_all_samplerates();

  135.     if (!formats)

  136.         return AVERROR(ENOMEM);

  137.     ff_set_common_samplerates(ctx, formats);

  138. 

  139.     return 0;

  140. }

  141. 

  142. static inline void scale_samples_u8(uint8_t *dst, const uint8_t *src,

  143.                                     int nb_samples, int volume)

  144. {

  145.     int i;

  146.     for (i = 0; i < nb_samples; i++)

  147.         dst[i] = av_clip_uint8(((((int64_t)src[i] - 128) * volume + 128) >> 8) + 128);

  148. }

  149. 

  150. static inline void scale_samples_u8_small(uint8_t *dst, const uint8_t *src,

  151.                                           int nb_samples, int volume)

  152. {

  153.     int i;

  154.     for (i = 0; i < nb_samples; i++)

  155.         dst[i] = av_clip_uint8((((src[i] - 128) * volume + 128) >> 8) + 128);

  156. }

  157. 

  158. static inline void scale_samples_s16(uint8_t *dst, const uint8_t *src,

  159.                                      int nb_samples, int volume)

  160. {

  161.     int i;

  162.     int16_t *smp_dst       = (int16_t *)dst;

  163.     const int16_t *smp_src = (const int16_t *)src;

  164.     for (i = 0; i < nb_samples; i++)

  165.         smp_dst[i] = av_clip_int16(((int64_t)smp_src[i] * volume + 128) >> 8);

  166. }

  167. 

  168. static inline void scale_samples_s16_small(uint8_t *dst, const uint8_t *src,

  169.                                            int nb_samples, int volume)

  170. {

  171.     int i;

  172.     int16_t *smp_dst       = (int16_t *)dst;

  173.     const int16_t *smp_src = (const int16_t *)src;

  174.     for (i = 0; i < nb_samples; i++)

  175.         smp_dst[i] = av_clip_int16((smp_src[i] * volume + 128) >> 8);

  176. }

  177. 

  178. static inline void scale_samples_s32(uint8_t *dst, const uint8_t *src,

  179.                                      int nb_samples, int volume)

  180. {

  181.     int i;

  182.     int32_t *smp_dst       = (int32_t *)dst;

  183.     const int32_t *smp_src = (const int32_t *)src;

  184.     for (i = 0; i < nb_samples; i++)

  185.         smp_dst[i] = av_clipl_int32((((int64_t)smp_src[i] * volume + 128) >> 8));

  186. }

  187. 

  188. 

  189. 

  190. static av_cold void volume_init(VolumeContext *vol)

  191. {

  192.     vol->samples_align = 1;

  193. 

  194.     switch (av_get_packed_sample_fmt(vol->sample_fmt)) {

  195.     case AV_SAMPLE_FMT_U8:

  196.         if (vol->volume_i < 0x1000000)

  197.             vol->scale_samples = scale_samples_u8_small;

  198.         else

  199.             vol->scale_samples = scale_samples_u8;

  200.         break;

  201.     case AV_SAMPLE_FMT_S16:

  202.         if (vol->volume_i < 0x10000)

  203.             vol->scale_samples = scale_samples_s16_small;

  204.         else

  205.             vol->scale_samples = scale_samples_s16;

  206.         break;

  207.     case AV_SAMPLE_FMT_S32:

  208.         vol->scale_samples = scale_samples_s32;

  209.         break;

  210.     case AV_SAMPLE_FMT_FLT:

  211.         avpriv_float_dsp_init(&vol->fdsp, 0);

  212.         vol->samples_align = 4;

  213.         break;

  214.     case AV_SAMPLE_FMT_DBL:

  215.         avpriv_float_dsp_init(&vol->fdsp, 0);

  216.         vol->samples_align = 8;

  217.         break;

  218.     }

  219. 

  220.     if (ARCH_X86)

  221.         ff_volume_init_x86(vol);

  222. }

  223. 

  224. static int config_output(AVFilterLink *outlink)

  225. {

  226.     AVFilterContext *ctx = outlink->src;

  227.     VolumeContext *vol   = ctx->priv;

  228.     AVFilterLink *inlink = ctx->inputs[0];

  229. 

  230.     vol->sample_fmt = inlink->format;

  231.     vol->channels   = av_get_channel_layout_nb_channels(inlink->channel_layout);

  232.     vol->planes     = av_sample_fmt_is_planar(inlink->format) ? vol->channels : 1;

  233. 

  234.     volume_init(vol);

  235. 

  236.     return 0;

  237. }

  238. 

  239. static int filter_frame(AVFilterLink *inlink, AVFrame *buf)

  240. {

  241.     VolumeContext *vol    = inlink->dst->priv;

  242.     AVFilterLink *outlink = inlink->dst->outputs[0];

  243.     int nb_samples        = buf->nb_samples;

  244.     AVFrame *out_buf;

  245.     AVFrameSideData *sd = av_frame_get_side_data(buf, AV_FRAME_DATA_REPLAYGAIN);

  246.     int ret;

  247. 

  248.     if (sd && vol->replaygain != REPLAYGAIN_IGNORE) {

  249.         if (vol->replaygain != REPLAYGAIN_DROP) {

  250.             AVReplayGain *replaygain = (AVReplayGain*)sd->data;

  251.             int32_t gain  = 100000;

  252.             uint32_t peak = 100000;

  253.             float g, p;

  254. 

  255.             if (vol->replaygain == REPLAYGAIN_TRACK &&

  256.                 replaygain->track_gain != INT32_MIN) {

  257.                 gain = replaygain->track_gain;

  258. 

  259.                 if (replaygain->track_peak != 0)

  260.                     peak = replaygain->track_peak;

  261.             } else if (replaygain->album_gain != INT32_MIN) {

  262.                 gain = replaygain->album_gain;

  263. 

  264.                 if (replaygain->album_peak != 0)

  265.                     peak = replaygain->album_peak;

  266.             } else {

  267.                 av_log(inlink->dst, AV_LOG_WARNING, "Both ReplayGain gain "

  268.                        "values are unknown.\n");

  269.             }

  270.             g = gain / 100000.0f;

  271.             p = peak / 100000.0f;

  272. 

  273.             av_log(inlink->dst, AV_LOG_VERBOSE,

  274.                    "Using gain %f dB from replaygain side data.\n", g);

  275. 

  276.             vol->volume   = pow(10, (g + vol->replaygain_preamp) / 20);

  277.             if (vol->replaygain_noclip)

  278.                 vol->volume = FFMIN(vol->volume, 1.0 / p);

  279.             vol->volume_i = (int)(vol->volume * 256 + 0.5);

  280. 

  281.             volume_init(vol);

  282.         }

  283.         av_frame_remove_side_data(buf, AV_FRAME_DATA_REPLAYGAIN);

  284.     }

  285. 

  286.     if (vol->volume == 1.0 || vol->volume_i == 256)

  287.         return ff_filter_frame(outlink, buf);

  288. 

  289.     /* do volume scaling in-place if input buffer is writable */

  290.     if (av_frame_is_writable(buf)) {

  291.         out_buf = buf;

  292.     } else {

  293.         out_buf = ff_get_audio_buffer(inlink, nb_samples);

  294.         if (!out_buf)

  295.             return AVERROR(ENOMEM);

  296.         ret = av_frame_copy_props(out_buf, buf);

  297.         if (ret < 0) {

  298.             av_frame_free(&out_buf);

  299.             av_frame_free(&buf);

  300.             return ret;

  301.         }

  302.     }

  303. 

  304.     if (vol->precision != PRECISION_FIXED || vol->volume_i > 0) {

  305.         int p, plane_samples;

  306. 

  307.         if (av_sample_fmt_is_planar(buf->format))

  308.             plane_samples = FFALIGN(nb_samples, vol->samples_align);

  309.         else

  310.             plane_samples = FFALIGN(nb_samples * vol->channels, vol->samples_align);

  311. 

  312.         if (vol->precision == PRECISION_FIXED) {

  313.             for (p = 0; p < vol->planes; p++) {

  314.                 vol->scale_samples(out_buf->extended_data[p],

  315.                                    buf->extended_data[p], plane_samples,

  316.                                    vol->volume_i);

  317.             }

  318.         } else if (av_get_packed_sample_fmt(vol->sample_fmt) == AV_SAMPLE_FMT_FLT) {

  319.             for (p = 0; p < vol->planes; p++) {

  320.                 vol->fdsp.vector_fmul_scalar((float *)out_buf->extended_data[p],

  321.                                              (const float *)buf->extended_data[p],

  322.                                              vol->volume, plane_samples);

  323.             }

  324.         } else {

  325.             for (p = 0; p < vol->planes; p++) {

  326.                 vol->fdsp.vector_dmul_scalar((double *)out_buf->extended_data[p],

  327.                                              (const double *)buf->extended_data[p],

  328.                                              vol->volume, plane_samples);

  329.             }

  330.         }

  331.     }

  332. 

  333.     emms_c();

  334. 

  335.     if (buf != out_buf)

  336.         av_frame_free(&buf);

  337. 

  338.     return ff_filter_frame(outlink, out_buf);

  339. }

  340. 

  341. static const AVFilterPad avfilter_af_volume_inputs[] = {

  342.     {

  343.         .name           = "default",

  344.         .type           = AVMEDIA_TYPE_AUDIO,

  345.         .filter_frame   = filter_frame,

  346.     },

  347.     { NULL }

  348. };

  349. 

  350. static const AVFilterPad avfilter_af_volume_outputs[] = {

  351.     {

  352.         .name         = "default",

  353.         .type         = AVMEDIA_TYPE_AUDIO,

  354.         .config_props = config_output,

  355.     },

  356.     { NULL }

  357. };

  358. 

  359. AVFilter ff_af_volume = {

  360.     .name           = "volume",

  361.     .description    = NULL_IF_CONFIG_SMALL("Change input volume."),

  362.     .query_formats  = query_formats,

  363.     .priv_size      = sizeof(VolumeContext),

  364.     .priv_class     = &volume_class,

  365.     .init           = init,

  366.     .inputs         = avfilter_af_volume_inputs,

  367.     .outputs        = avfilter_af_volume_outputs,

  368. };