]> git.sesse.net Git - ffmpeg/blobdiff - libavcodec/aacenc.c
Reindent after last commit.
[ffmpeg] / libavcodec / aacenc.c
index 71aa0e37a574c01db1eb5f630a3697507537364a..0de6622389ef90eb7fbad67f32a584ced63fec70 100644 (file)
@@ -2,20 +2,20 @@
  * AAC encoder
  * Copyright (C) 2008 Konstantin Shishkov
  *
- * This file is part of Libav.
+ * This file is part of FFmpeg.
  *
- * Libav is free software; you can redistribute it and/or
+ * FFmpeg is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public
  * License as published by the Free Software Foundation; either
  * version 2.1 of the License, or (at your option) any later version.
  *
- * Libav is distributed in the hope that it will be useful,
+ * FFmpeg is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  * Lesser General Public License for more details.
  *
  * You should have received a copy of the GNU Lesser General Public
- * License along with Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
  * add temporal noise shaping
  ***********************************/
 
+#include "libavutil/opt.h"
 #include "avcodec.h"
 #include "put_bits.h"
 #include "dsputil.h"
 #include "mpeg4audio.h"
+#include "kbdwin.h"
+#include "sinewin.h"
 
 #include "aac.h"
 #include "aactab.h"
@@ -132,6 +135,15 @@ static const uint8_t aac_chan_configs[6][5] = {
  {4, TYPE_SCE, TYPE_CPE, TYPE_CPE, TYPE_LFE}, // 6 channels - front center + stereo + back stereo + LFE
 };
 
+static const uint8_t channel_maps[][AAC_MAX_CHANNELS] = {
+    { 0 },
+    { 0, 1 },
+    { 2, 0, 1 },
+    { 2, 0, 1, 3 },
+    { 2, 0, 1, 3, 4 },
+    { 2, 0, 1, 4, 5, 3 },
+};
+
 /**
  * Make AAC audio config object.
  * @see 1.6.2.1 "Syntax - AudioSpecificConfig"
@@ -250,7 +262,7 @@ static void apply_window_and_mdct(AVCodecContext *avctx, AACEncContext *s,
             for (i = 0; i < 1024; i++)
                 sce->saved[i] = audio[i * chans];
         }
-        ff_mdct_calc(&s->mdct1024, sce->coeffs, output);
+        s->mdct1024.mdct_calc(&s->mdct1024, sce->coeffs, output);
     } else {
         for (k = 0; k < 1024; k += 128) {
             for (i = 448 + k; i < 448 + k + 256; i++)
@@ -259,7 +271,7 @@ static void apply_window_and_mdct(AVCodecContext *avctx, AACEncContext *s,
                                          : audio[(i-1024)*chans];
             s->dsp.vector_fmul        (output,     output, k ?  swindow : pwindow, 128);
             s->dsp.vector_fmul_reverse(output+128, output+128, swindow, 128);
-            ff_mdct_calc(&s->mdct128, sce->coeffs + k, output);
+            s->mdct128.mdct_calc(&s->mdct128, sce->coeffs + k, output);
         }
         for (i = 0; i < 1024; i++)
             sce->saved[i] = audio[i * chans];
@@ -487,7 +499,7 @@ static int aac_encode_frame(AVCodecContext *avctx,
     AACEncContext *s = avctx->priv_data;
     int16_t *samples = s->samples, *samples2, *la;
     ChannelElement *cpe;
-    int i, j, chans, tag, start_ch;
+    int i, ch, w, g, chans, tag, start_ch;
     const uint8_t *chan_map = aac_chan_configs[avctx->channels-1];
     int chan_el_counter[4];
     FFPsyWindowInfo windows[AAC_MAX_CHANNELS];
@@ -496,15 +508,24 @@ static int aac_encode_frame(AVCodecContext *avctx,
         return 0;
     if (data) {
         if (!s->psypp) {
-            memcpy(s->samples + 1024 * avctx->channels, data,
-                   1024 * avctx->channels * sizeof(s->samples[0]));
+            if (avctx->channels <= 2) {
+                memcpy(s->samples + 1024 * avctx->channels, data,
+                       1024 * avctx->channels * sizeof(s->samples[0]));
+            } else {
+                for (i = 0; i < 1024; i++)
+                    for (ch = 0; ch < avctx->channels; ch++)
+                        s->samples[(i + 1024) * avctx->channels + ch] =
+                            ((int16_t*)data)[i * avctx->channels +
+                                             channel_maps[avctx->channels-1][ch]];
+            }
         } else {
             start_ch = 0;
             samples2 = s->samples + 1024 * avctx->channels;
             for (i = 0; i < chan_map[0]; i++) {
                 tag = chan_map[i+1];
                 chans = tag == TYPE_CPE ? 2 : 1;
-                ff_psy_preprocess(s->psypp, (uint16_t*)data + start_ch,
+                ff_psy_preprocess(s->psypp,
+                                  (uint16_t*)data + channel_maps[avctx->channels-1][start_ch],
                                   samples2 + start_ch, start_ch, chans);
                 start_ch += chans;
             }
@@ -522,34 +543,33 @@ static int aac_encode_frame(AVCodecContext *avctx,
         tag      = chan_map[i+1];
         chans    = tag == TYPE_CPE ? 2 : 1;
         cpe      = &s->cpe[i];
-        for (j = 0; j < chans; j++) {
-            IndividualChannelStream *ics = &cpe->ch[j].ics;
-            int k;
-            int cur_channel = start_ch + j;
+        for (ch = 0; ch < chans; ch++) {
+            IndividualChannelStream *ics = &cpe->ch[ch].ics;
+            int cur_channel = start_ch + ch;
             samples2 = samples + cur_channel;
             la       = samples2 + (448+64) * avctx->channels;
             if (!data)
                 la = NULL;
             if (tag == TYPE_LFE) {
-                wi[j].window_type[0] = ONLY_LONG_SEQUENCE;
-                wi[j].window_shape   = 0;
-                wi[j].num_windows    = 1;
-                wi[j].grouping[0]    = 1;
+                wi[ch].window_type[0] = ONLY_LONG_SEQUENCE;
+                wi[ch].window_shape   = 0;
+                wi[ch].num_windows    = 1;
+                wi[ch].grouping[0]    = 1;
             } else {
-                wi[j] = ff_psy_suggest_window(&s->psy, samples2, la, cur_channel,
+                wi[ch] = s->psy.model->window(&s->psy, samples2, la, cur_channel,
                                               ics->window_sequence[0]);
             }
             ics->window_sequence[1] = ics->window_sequence[0];
-            ics->window_sequence[0] = wi[j].window_type[0];
+            ics->window_sequence[0] = wi[ch].window_type[0];
             ics->use_kb_window[1]   = ics->use_kb_window[0];
-            ics->use_kb_window[0]   = wi[j].window_shape;
-            ics->num_windows        = wi[j].num_windows;
+            ics->use_kb_window[0]   = wi[ch].window_shape;
+            ics->num_windows        = wi[ch].num_windows;
             ics->swb_sizes          = s->psy.bands    [ics->num_windows == 8];
             ics->num_swb            = tag == TYPE_LFE ? 12 : s->psy.num_bands[ics->num_windows == 8];
-            for (k = 0; k < ics->num_windows; k++)
-                ics->group_len[k] = wi[j].grouping[k];
+            for (w = 0; w < ics->num_windows; w++)
+                ics->group_len[w] = wi[ch].grouping[w];
 
-            apply_window_and_mdct(avctx, s, &cpe->ch[j], samples2);
+            apply_window_and_mdct(avctx, s, &cpe->ch[ch], samples2);
         }
         start_ch += chans;
     }
@@ -567,10 +587,10 @@ static int aac_encode_frame(AVCodecContext *avctx,
             cpe      = &s->cpe[i];
             put_bits(&s->pb, 3, tag);
             put_bits(&s->pb, 4, chan_el_counter[tag]++);
-            for (j = 0; j < chans; j++) {
-                s->cur_channel = start_ch + j;
-                ff_psy_set_band_info(&s->psy, s->cur_channel, cpe->ch[j].coeffs, &wi[j]);
-                s->coder->search_for_quantizers(avctx, s, &cpe->ch[j], s->lambda);
+            for (ch = 0; ch < chans; ch++) {
+                s->cur_channel = start_ch + ch;
+                s->psy.model->analyze(&s->psy, s->cur_channel, cpe->ch[ch].coeffs, &wi[ch]);
+                s->coder->search_for_quantizers(avctx, s, &cpe->ch[ch], s->lambda);
             }
             cpe->common_window = 0;
             if (chans > 1
@@ -578,16 +598,24 @@ static int aac_encode_frame(AVCodecContext *avctx,
                 && wi[0].window_shape   == wi[1].window_shape) {
 
                 cpe->common_window = 1;
-                for (j = 0; j < wi[0].num_windows; j++) {
-                    if (wi[0].grouping[j] != wi[1].grouping[j]) {
+                for (w = 0; w < wi[0].num_windows; w++) {
+                    if (wi[0].grouping[w] != wi[1].grouping[w]) {
                         cpe->common_window = 0;
                         break;
                     }
                 }
             }
             s->cur_channel = start_ch;
-            if (cpe->common_window && s->coder->search_for_ms)
-                s->coder->search_for_ms(s, cpe, s->lambda);
+            if (s->options.stereo_mode && cpe->common_window) {
+                if (s->options.stereo_mode > 0) {
+                    IndividualChannelStream *ics = &cpe->ch[0].ics;
+                    for (w = 0; w < ics->num_windows; w += ics->group_len[w])
+                        for (g = 0;  g < ics->num_swb; g++)
+                            cpe->ms_mask[w*16+g] = 1;
+                } else if (s->coder->search_for_ms) {
+                    s->coder->search_for_ms(s, cpe, s->lambda);
+                }
+            }
             adjust_frame_information(s, cpe, chans);
             if (chans == 2) {
                 put_bits(&s->pb, 1, cpe->common_window);
@@ -596,16 +624,18 @@ static int aac_encode_frame(AVCodecContext *avctx,
                     encode_ms_info(&s->pb, cpe);
                 }
             }
-            for (j = 0; j < chans; j++) {
-                s->cur_channel = start_ch + j;
-                encode_individual_channel(avctx, s, &cpe->ch[j], cpe->common_window);
+            for (ch = 0; ch < chans; ch++) {
+                s->cur_channel = start_ch + ch;
+                encode_individual_channel(avctx, s, &cpe->ch[ch], cpe->common_window);
             }
             start_ch += chans;
         }
 
         frame_bits = put_bits_count(&s->pb);
-        if (frame_bits <= 6144 * avctx->channels - 3)
+        if (frame_bits <= 6144 * avctx->channels - 3) {
+            s->psy.bitres.bits = frame_bits / avctx->channels;
             break;
+        }
 
         s->lambda *= avctx->bit_rate * 1024.0f / avctx->sample_rate / frame_bits;
 
@@ -642,6 +672,22 @@ static av_cold int aac_encode_end(AVCodecContext *avctx)
     return 0;
 }
 
+#define AACENC_FLAGS AV_OPT_FLAG_ENCODING_PARAM | AV_OPT_FLAG_AUDIO_PARAM
+static const AVOption aacenc_options[] = {
+    {"stereo_mode", "Stereo coding method", offsetof(AACEncContext, options.stereo_mode), FF_OPT_TYPE_INT, {.dbl = 0}, -1, 1, AACENC_FLAGS, "stereo_mode"},
+        {"auto",     "Selected by the Encoder", 0, FF_OPT_TYPE_CONST, {.dbl = -1 }, INT_MIN, INT_MAX, AACENC_FLAGS, "stereo_mode"},
+        {"ms_off",   "Disable Mid/Side coding", 0, FF_OPT_TYPE_CONST, {.dbl =  0 }, INT_MIN, INT_MAX, AACENC_FLAGS, "stereo_mode"},
+        {"ms_force", "Force Mid/Side for the whole frame if possible", 0, FF_OPT_TYPE_CONST, {.dbl =  1 }, INT_MIN, INT_MAX, AACENC_FLAGS, "stereo_mode"},
+    {NULL}
+};
+
+static const AVClass aacenc_class = {
+    "AAC encoder",
+    av_default_item_name,
+    aacenc_options,
+    LIBAVUTIL_VERSION_INT,
+};
+
 AVCodec ff_aac_encoder = {
     "aac",
     AVMEDIA_TYPE_AUDIO,
@@ -653,4 +699,5 @@ AVCodec ff_aac_encoder = {
     .capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY | CODEC_CAP_EXPERIMENTAL,
     .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
     .long_name = NULL_IF_CONFIG_SMALL("Advanced Audio Coding"),
+    .priv_class = &aacenc_class,
 };