1 /*****************************************************************************
2 * aac.c: FDK-AAC Encoder plugin for vlc.
3 *****************************************************************************
4 * Copyright (C) 2012 Sergio Ammirata
6 * Authors: Sergio Ammirata <sergio@ammirata.net>
8 * This library is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * This library is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with this library; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 * Alternatively you can redistribute this file under the terms of the
23 * BSD license as stated below:
25 * Redistribution and use in source and binary forms, with or without
26 * modification, are permitted provided that the following conditions
28 * 1. Redistributions of source code must retain the above copyright
29 * notice, this list of conditions and the following disclaimer.
30 * 2. Redistributions in binary form must reproduce the above copyright
31 * notice, this list of conditions and the following disclaimer in
32 * the documentation and/or other materials provided with the
35 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
36 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
37 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
38 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
39 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
40 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
41 * TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
42 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
43 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
44 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
45 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
47 *****************************************************************************/
49 /*****************************************************************************
51 *****************************************************************************/
56 #include <fdk-aac/aacenc_lib.h>
58 #include <vlc_common.h>
59 #include <vlc_plugin.h>
60 #include <vlc_codec.h>
62 static int OpenEncoder( vlc_object_t * );
63 static void CloseEncoder( vlc_object_t * );
65 #define ENC_CFG_PREFIX "sout-fdkaac-"
67 #define AOT_TEXT N_("Encoder Profile")
68 #define AOT_LONGTEXT N_( "Encoder Algorithm to use" )
70 #define SIDEBAND_TEXT N_("Enable spectral band replication")
71 #define SIDEBAND_LONGTEXT N_( "This is an optional feature only for the AAC-ELD profile" )
73 #define VBR_QUALITY_TEXT N_("VBR Quality")
74 #define VBR_QUALITY_LONGTEXT N_( "Quality of the VBR Encoding (0=cbr, 1-5 constant quality vbr, 5 is best" )
76 #define AFTERBURNER_TEXT N_("Enable afterburner library")
77 #define AFTERBURNER_LONGTEXT N_( "This library will produce higher quality audio at the expense of additional CPU usage (default is enabled)" )
79 #define SIGNALING_TEXT N_("Signaling mode of the extension AOT")
80 #define SIGNALING_LONGTEXT N_( "1 is explicit for SBR and implicit for PS (default), 2 is explicit hierarchical" )
82 #define CH_ORDER_MPEG 0 /*!< MPEG channel ordering (e. g. 5.1: C, L, R, SL, SR, LFE) */
83 #define CH_ORDER_WAV 1 /*!< WAV fileformat channel ordering (e. g. 5.1: L, R, C, LFE, SL, SR) */
84 #define CH_ORDER_WG4 2 /*!< WG4 fileformat channel ordering (e. g. 5.1: L, R, SL, SR, C, LFE) */
86 #define PROFILE_AAC_LC 2
87 #define PROFILE_AAC_HE 5
88 #define PROFILE_AAC_HE_v2 29
89 #define PROFILE_AAC_LD 23
90 #define PROFILE_AAC_ELD 39
92 #define SIGNALING_COMPATIBLE 1
93 #define SIGNALING_HIERARCHICAL 2
95 static const int pi_aot_values[] = { PROFILE_AAC_LC, PROFILE_AAC_HE, PROFILE_AAC_HE_v2, PROFILE_AAC_LD, PROFILE_AAC_ELD };
96 static const char *const ppsz_aot_descriptions[] =
97 { N_("AAC-LC"), N_("HE-AAC"), N_("HE-AAC-v2"), N_("AAC-LD"), N_("AAC-ELD") };
100 set_shortname( N_("FDKAAC") )
101 set_description( N_("FDK-AAC Audio encoder") )
102 set_capability( "encoder", 150 )
103 set_callbacks( OpenEncoder, CloseEncoder )
104 add_shortcut( "fdkaac" )
105 set_category( CAT_INPUT )
106 set_subcategory( SUBCAT_INPUT_ACODEC )
107 add_integer( ENC_CFG_PREFIX "profile", PROFILE_AAC_LC, AOT_TEXT,
108 AOT_LONGTEXT, false )
109 change_integer_list( pi_aot_values, ppsz_aot_descriptions );
110 add_bool( ENC_CFG_PREFIX "sbr", false, SIDEBAND_TEXT,
111 SIDEBAND_LONGTEXT, false )
112 add_integer( ENC_CFG_PREFIX "vbr", 0, VBR_QUALITY_TEXT,
113 VBR_QUALITY_LONGTEXT, false )
114 change_integer_range (0, 5)
115 add_bool( ENC_CFG_PREFIX "afterburner", true, AFTERBURNER_TEXT,
116 AFTERBURNER_LONGTEXT, true )
117 add_integer( ENC_CFG_PREFIX "signaling", SIGNALING_COMPATIBLE, SIGNALING_TEXT,
118 SIGNALING_LONGTEXT, true )
119 change_integer_range (0, 2)
122 /*****************************************************************************
124 *****************************************************************************/
125 static block_t *EncodeAudio( encoder_t *p_enc, block_t *p_buf );
127 static const char *const ppsz_enc_options[] = {
128 "profile", "sbr", "vbr", "afterburner", "signaling", NULL
131 /*****************************************************************************
132 * encoder_sys_t : aac encoder descriptor
133 *****************************************************************************/
136 double d_compression_ratio;
138 int i_aot; /* This stores the aac profile chosen */
139 int i_vbr; /* cbr or vbr-quality value chosen */
140 int i_signaling; /* Library feature for backwards compatibility */
141 int i_encoderdelay; /* Samples delay introduced by the profile */
143 int i_maxoutputsize; /* Maximum buffer size for encoded output */
144 HANDLE_AACENCODER handle;
145 bool b_afterburner; /* Library feature for additional quality */
146 bool b_eld_sbr; /* Spectral band replication option for ELD profile */
149 static const char *aac_get_errorstring(AACENC_ERROR erraac)
154 case AACENC_INVALID_HANDLE:
155 return "Invalid handle";
156 case AACENC_MEMORY_ERROR:
157 return "Memory allocation error";
158 case AACENC_UNSUPPORTED_PARAMETER:
159 return "Unsupported parameter";
160 case AACENC_INVALID_CONFIG:
161 return "Invalid config";
162 case AACENC_INIT_ERROR:
163 return "Initialization error";
164 case AACENC_INIT_AAC_ERROR:
165 return "AAC library initialization error";
166 case AACENC_INIT_SBR_ERROR:
167 return "SBR library initialization error";
168 case AACENC_INIT_TP_ERROR:
169 return "Transport library initialization error";
170 case AACENC_INIT_META_ERROR:
171 return "Metadata library initialization error";
172 case AACENC_ENCODE_ERROR:
173 return "Encoding error";
174 case AACENC_ENCODE_EOF:
175 return "End of file";
177 return "Unknown error";
181 /*****************************************************************************
182 * OpenDecoder: open the encoder.
183 *****************************************************************************/
184 static int OpenEncoder( vlc_object_t *p_this )
187 encoder_sys_t *p_sys;
190 bool b_profile_selected;
196 p_enc = (encoder_t *)p_this;
197 b_profile_selected = false;
201 if( p_enc->fmt_out.i_codec != VLC_FOURCC( 'l', 'a', 'a', 'c' ) &&
202 p_enc->fmt_out.i_codec != VLC_FOURCC( 'h', 'a', 'a', 'c' ) &&
203 p_enc->fmt_out.i_codec != VLC_FOURCC( 's', 'a', 'a', 'c' ) &&
204 p_enc->fmt_out.i_codec != VLC_CODEC_MP4A )
208 else if ( p_enc->fmt_out.i_codec == VLC_FOURCC( 'l', 'a', 'a', 'c' ) )
210 b_profile_selected = true;
211 i_profile = PROFILE_AAC_LC;
213 else if ( p_enc->fmt_out.i_codec == VLC_FOURCC( 'h', 'a', 'a', 'c' ) )
215 b_profile_selected = true;
216 i_profile = PROFILE_AAC_HE;
218 else if ( p_enc->fmt_out.i_codec == VLC_FOURCC( 's', 'a', 'a', 'c' ) )
220 b_profile_selected = true;
221 i_profile = PROFILE_AAC_HE_v2;
224 switch (p_enc->fmt_in.audio.i_channels) {
225 case 1: mode = MODE_1; sce = 1; cpe = 0; break;
226 case 2: mode = MODE_2; sce = 0; cpe = 1; break;
227 case 3: mode = MODE_1_2; sce = 1; cpe = 1; break;
228 case 4: mode = MODE_1_2_1; sce = 2; cpe = 1; break;
229 case 5: mode = MODE_1_2_2; sce = 1; cpe = 2; break;
230 case 6: mode = MODE_1_2_2_1; sce = 2; cpe = 2; break;
231 case 8: mode = MODE_1_2_2_2_1; sce = 2; cpe = 3; break;
233 msg_Err( p_enc, "we do not support > 8 input channels, this input has %i",
234 p_enc->fmt_in.audio.i_channels );
238 msg_Info(p_enc, "Initializing AAC Encoder, %i channels", p_enc->fmt_in.audio.i_channels);
240 /* Allocate the memory needed to store the encoder's structure */
241 p_sys = (encoder_sys_t *)malloc(sizeof(encoder_sys_t));
242 if( unlikely( !p_sys ) )
244 p_enc->p_sys = p_sys;
245 p_enc->fmt_in.i_codec = VLC_CODEC_S16N;
246 p_enc->fmt_out.i_cat = AUDIO_ES;
247 p_enc->fmt_out.i_codec = VLC_CODEC_MP4A;
249 config_ChainParse( p_enc, ENC_CFG_PREFIX, ppsz_enc_options, p_enc->p_cfg );
251 if ( b_profile_selected == false )
252 p_sys->i_aot = var_InheritInteger( p_enc, ENC_CFG_PREFIX "profile" );
254 p_sys->i_aot = i_profile;
255 p_sys->b_eld_sbr = var_InheritBool( p_enc, ENC_CFG_PREFIX "sbr" );
256 p_sys->i_vbr = var_InheritInteger( p_enc, ENC_CFG_PREFIX "vbr" );
257 p_sys->b_afterburner = var_InheritBool( p_enc, ENC_CFG_PREFIX "afterburner" );
258 p_sys->i_signaling = var_InheritInteger( p_enc, ENC_CFG_PREFIX "signaling" );
259 p_sys->i_pts_last = 0;
261 if ((p_sys->i_aot == PROFILE_AAC_HE || p_sys->i_aot == PROFILE_AAC_HE_v2) && p_sys->i_vbr > 3)
263 msg_Warn(p_enc, "Maximum VBR quality for this profile is 3, setting vbr=3");
266 if ((erraac = aacEncOpen(&p_sys->handle, 0, p_enc->fmt_in.audio.i_channels)) != AACENC_OK) {
267 msg_Err(p_enc, "Unable to open encoder: %s", aac_get_errorstring(erraac));
271 if ( p_sys->i_aot == PROFILE_AAC_HE_v2 && p_enc->fmt_in.audio.i_channels != 2 )
273 msg_Err(p_enc, "The HE-AAC-v2 profile can only be used with stereo sources");
276 if ( p_sys->i_aot == PROFILE_AAC_ELD && p_enc->fmt_in.audio.i_channels != 2 )
278 msg_Err(p_enc, "The ELD-AAC profile can only be used with stereo sources");
281 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_AOT, p_sys->i_aot)) != AACENC_OK) {
282 msg_Err(p_enc, "Unable to set the Profile %i: %s", p_sys->i_aot, aac_get_errorstring(erraac));
285 if (p_sys->i_aot == PROFILE_AAC_ELD && p_sys->b_eld_sbr) {
286 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_SBR_MODE, 1)) != AACENC_OK) {
287 msg_Err(p_enc, "Unable to set SBR mode for ELD: %s", aac_get_errorstring(erraac));
291 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_SAMPLERATE,
292 p_enc->fmt_out.audio.i_rate)) != AACENC_OK) {
293 msg_Err(p_enc, "Unable to set the sample rate %i: %s",p_enc->fmt_out.audio.i_rate,
294 aac_get_errorstring(erraac));
297 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_CHANNELMODE, mode)) != AACENC_OK) {
298 msg_Err(p_enc, "Unable to set the channel mode: %s", aac_get_errorstring(erraac));
301 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_CHANNELORDER, CH_ORDER_WG4)) != AACENC_OK) {
302 msg_Err(p_enc, "Unable to set the sound channel order: %s", aac_get_errorstring(erraac));
305 if (p_sys->i_vbr != 0) {
306 if ((erraac = aacEncoder_SetParam(p_sys->handle,
307 AACENC_BITRATEMODE, p_sys->i_vbr)) != AACENC_OK) {
308 msg_Err(p_enc, "Unable to set the VBR bitrate mode: %s", aac_get_errorstring(erraac));
312 if (p_enc->fmt_out.i_bitrate == 0) {
313 if (p_sys->i_aot == PROFILE_AAC_HE_v2) {
317 i_bitrate = (96*sce + 128*cpe) * p_enc->fmt_out.audio.i_rate / 44;
318 if (p_sys->i_aot == PROFILE_AAC_HE ||
319 p_sys->i_aot == PROFILE_AAC_HE_v2 ||
322 p_enc->fmt_out.i_bitrate = i_bitrate;
323 msg_Info(p_enc, "Setting optimal bitrate of %i", i_bitrate);
327 i_bitrate = p_enc->fmt_out.i_bitrate;
329 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_BITRATE,
330 i_bitrate)) != AACENC_OK) {
331 msg_Err(p_enc, "Unable to set the bitrate %i: %s", i_bitrate,
332 aac_get_errorstring(erraac));
336 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_TRANSMUX, 0)) != AACENC_OK) {
337 msg_Err(p_enc, "Unable to set the ADTS transmux: %s", aac_get_errorstring(erraac));
340 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_SIGNALING_MODE,
341 (int)p_sys->i_signaling)) != AACENC_OK) {
342 /* use explicit backward compatible =1 */
343 /* use explicit hierarchical signaling =2 */
344 msg_Err(p_enc, "Unable to set signaling mode: %s", aac_get_errorstring(erraac));
347 if ((erraac = aacEncoder_SetParam(p_sys->handle, AACENC_AFTERBURNER,
348 (int)p_sys->b_afterburner)) != AACENC_OK) {
349 msg_Err(p_enc, "Unable to set the afterburner mode: %s", aac_get_errorstring(erraac));
352 if ((erraac = aacEncEncode(p_sys->handle, NULL, NULL, NULL, NULL)) != AACENC_OK) {
353 msg_Err(p_enc, "Unable to initialize the encoder: %s", aac_get_errorstring(erraac));
356 AACENC_InfoStruct info = { 0 };
357 if ((erraac = aacEncInfo(p_sys->handle, &info)) != AACENC_OK) {
358 msg_Err(p_enc, "Unable to get the encoder info: %s", aac_get_errorstring(erraac));
362 /* The maximum packet size is 6144 bits aka 768 bytes per channel. */
363 p_sys->i_maxoutputsize = 768*p_enc->fmt_in.audio.i_channels;
364 p_enc->fmt_in.audio.i_bitspersample = 16;
365 p_sys->i_frame_size = info.frameLength;
366 p_sys->i_encoderdelay = info.encoderDelay;
368 p_enc->fmt_out.i_extra = info.confSize;
369 if( p_enc->fmt_out.i_extra )
371 p_enc->fmt_out.p_extra = malloc( p_enc->fmt_out.i_extra );
372 if ( p_enc->fmt_out.p_extra == NULL )
374 msg_Err(p_enc, "Unable to allocate fmt_out.p_extra");
377 memcpy( p_enc->fmt_out.p_extra, info.confBuf,
378 p_enc->fmt_out.i_extra );
381 p_enc->pf_encode_audio = EncodeAudio;
384 // TODO: Add more debug info to this config printout
385 msg_Dbg(p_enc, "fmt_out.p_extra = %i", p_enc->fmt_out.i_extra);
391 aacEncClose(&p_sys->handle);
396 /****************************************************************************
397 * EncodeAudio: the whole thing
398 ****************************************************************************/
399 static block_t *EncodeAudio( encoder_t *p_enc, block_t *p_aout_buf )
401 encoder_sys_t *p_sys;
410 p_sys = p_enc->p_sys;
413 if ( likely( p_aout_buf ) )
415 p_buffer = (int16_t *)p_aout_buf->p_buffer;
416 i_samples = p_aout_buf->i_nb_samples;
417 i_pts_out = p_aout_buf->i_pts - (mtime_t)((double)CLOCK_FREQ *
418 (double)p_sys->i_encoderdelay /
419 (double)p_enc->fmt_out.audio.i_rate);
420 if (p_sys->i_pts_last == 0)
421 p_sys->i_pts_last = i_pts_out - (mtime_t)((double)CLOCK_FREQ *
422 (double)(p_sys->i_frame_size) /
423 (double)p_enc->fmt_out.audio.i_rate);
428 i_pts_out = p_sys->i_pts_last;
431 i_samples_left = i_samples;
434 while ( i_samples_left >= 0 )
436 AACENC_BufDesc in_buf = { 0 }, out_buf = { 0 };
437 AACENC_InArgs in_args = { 0 };
438 AACENC_OutArgs out_args = { 0 };
439 int in_identifier = IN_AUDIO_DATA;
440 int in_size, in_elem_size;
441 int out_identifier = OUT_BITSTREAM_DATA;
442 int out_size, out_elem_size;
443 void *in_ptr, *out_ptr;
445 if ( unlikely(i_samples == 0) ) {
446 // this forces the encoder to purge whatever is left in the internal buffer
447 in_args.numInSamples = -1;
449 in_ptr = p_buffer + (i_samples - i_samples_left)*p_enc->fmt_in.audio.i_channels;
450 in_size = 2*p_enc->fmt_in.audio.i_channels*i_samples_left;
452 in_args.numInSamples = p_enc->fmt_in.audio.i_channels*i_samples_left;
454 in_buf.bufs = &in_ptr;
455 in_buf.bufferIdentifiers = &in_identifier;
456 in_buf.bufSizes = &in_size;
457 in_buf.bufElSizes = &in_elem_size;
460 p_block = block_Alloc( p_sys->i_maxoutputsize );
461 p_block->i_buffer = p_sys->i_maxoutputsize;
462 out_ptr = p_block->p_buffer;
463 out_size = p_block->i_buffer;
466 out_buf.bufs = &out_ptr;
467 out_buf.bufferIdentifiers = &out_identifier;
468 out_buf.bufSizes = &out_size;
469 out_buf.bufElSizes = &out_elem_size;
471 if ((erraac = aacEncEncode(p_sys->handle, &in_buf, &out_buf, &in_args, &out_args)) != AACENC_OK) {
472 if (erraac == AACENC_ENCODE_EOF) {
473 msg_Info( p_enc, "Encoding final bytes (EOF)");
477 msg_Err( p_enc, "Encoding failed: %s", aac_get_errorstring(erraac));
478 block_Release(p_block);
482 if ( out_args.numOutBytes > 0 )
484 p_block->i_buffer = out_args.numOutBytes;
485 if ( unlikely(i_samples == 0) )
487 // I only have the numOutBytes so approximate based on compression factor
488 double d_samples_forward = p_sys->d_compression_ratio*(double)out_args.numOutBytes;
489 i_pts_out += (mtime_t)d_samples_forward;
490 p_block->i_length = (mtime_t)d_samples_forward;
491 // TODO: It would be more precise (a few microseconds) to use d_samples_forward =
492 // (mtime_t)CLOCK_FREQ * (mtime_t)p_sys->i_frame_size/(mtime_t)p_enc->fmt_out.audio.i_rate
493 // but I am not sure if the lib always outputs a full frame when
494 // emptying the internal buffer in the EOF scenario
498 if ( i_loop_count == 0 )
500 // There can be an implicit delay in the first loop cycle because leftover bytes
501 // in the library buffer from the prior block
502 double d_samples_delay = (double)p_sys->i_frame_size - (double)out_args.numInSamples /
503 (double)p_enc->fmt_in.audio.i_channels;
504 i_pts_out -= (mtime_t)((double)CLOCK_FREQ * d_samples_delay /
505 (double)p_enc->fmt_out.audio.i_rate);
506 p_block->i_length = (mtime_t)((double)CLOCK_FREQ * (double)p_sys->i_frame_size /
507 (double)p_enc->fmt_out.audio.i_rate);
508 p_block->i_nb_samples = d_samples_delay;
509 //p_block->i_length = i_pts_out - p_sys->i_pts_last;
513 double d_samples_forward = (double)out_args.numInSamples/(double)p_enc->fmt_in.audio.i_channels;
514 double d_length = ((double)CLOCK_FREQ * d_samples_forward /
515 (double)p_enc->fmt_out.audio.i_rate);
516 i_pts_out += (mtime_t) d_length;
517 p_block->i_length = (mtime_t) d_length;
518 p_block->i_nb_samples = d_samples_forward;
521 p_block->i_dts = p_block->i_pts = i_pts_out;
522 block_ChainAppend( &p_chain, p_block );
523 //msg_Dbg( p_enc, "p_block->i_dts %llu, p_block->i_length %llu, p_sys->i_pts_last %llu\n\
524 // out_args.numOutBytes = %i, out_args.numInSamples = %i, i_samples %i, i_loop_count %i",
525 // p_block->i_dts , p_block->i_length,p_sys->i_pts_last,
526 // out_args.numOutBytes, out_args.numInSamples, i_samples, i_loop_count);
527 if ( likely(i_samples > 0) )
529 p_sys->d_compression_ratio = (double)p_block->i_length / (double)out_args.numOutBytes;
530 i_samples_left -= out_args.numInSamples/p_enc->fmt_in.audio.i_channels;
531 p_sys->i_pts_last = i_pts_out;
536 block_Release(p_block);
537 //msg_Dbg( p_enc, "aac_encode_audio: not enough data yet");
540 if ( unlikely(i_loop_count++ > 100) )
542 msg_Err( p_enc, "Loop count greater than 100!!!, something must be wrong with the encoder library");
551 /*****************************************************************************
552 * CloseDecoder: decoder destruction
553 *****************************************************************************/
554 static void CloseEncoder( vlc_object_t *p_this )
556 encoder_t *p_enc = (encoder_t *)p_this;
557 encoder_sys_t *p_sys = p_enc->p_sys;
559 aacEncClose(&p_sys->handle);