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,
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_encoderdelay; /* Samples delay introduced by the profile */
140 int i_maxoutputsize; /* Maximum buffer size for encoded output */
141 HANDLE_AACENCODER handle;
144 static const char *fdkaac_error(AACENC_ERROR erraac)
147 case AACENC_OK: return "No error";
148 case AACENC_INVALID_HANDLE: return "Invalid handle";
149 case AACENC_MEMORY_ERROR: return "Memory allocation error";
150 case AACENC_UNSUPPORTED_PARAMETER: return "Unsupported parameter";
151 case AACENC_INVALID_CONFIG: return "Invalid config";
152 case AACENC_INIT_ERROR: return "Initialization error";
153 case AACENC_INIT_AAC_ERROR: return "AAC library initialization error";
154 case AACENC_INIT_SBR_ERROR: return "SBR library initialization error";
155 case AACENC_INIT_TP_ERROR: return "Transport library initialization error";
156 case AACENC_INIT_META_ERROR: return "Metadata library initialization error";
157 case AACENC_ENCODE_ERROR: return "Encoding error";
158 case AACENC_ENCODE_EOF: return "End of file";
159 default: return "Unknown error";
163 /*****************************************************************************
164 * OpenDecoder: open the encoder.
165 *****************************************************************************/
166 static int OpenEncoder(vlc_object_t *p_this)
168 encoder_t *p_enc = (encoder_t *)p_this;
170 config_ChainParse(p_enc, ENC_CFG_PREFIX, ppsz_enc_options, p_enc->p_cfg);
173 switch (p_enc->fmt_out.i_codec) {
175 i_aot = var_InheritInteger(p_enc, ENC_CFG_PREFIX "profile");
177 case VLC_FOURCC('l', 'a', 'a', 'c'):
178 i_aot = PROFILE_AAC_LC;
180 case VLC_FOURCC('h', 'a', 'a', 'c'):
181 i_aot = PROFILE_AAC_HE;
183 case VLC_FOURCC('s', 'a', 'a', 'c'):
184 i_aot = PROFILE_AAC_HE_v2;
190 if (p_enc->fmt_in.audio.i_channels != 2)
191 if (i_aot == PROFILE_AAC_HE_v2 || i_aot == PROFILE_AAC_ELD) {
192 msg_Err(p_enc, "Selected profile %d can only be used with stereo", i_aot);
196 uint16_t channel_config;
198 switch (p_enc->fmt_in.audio.i_channels) {
199 case 1: mode = MODE_1; channel_config = AOUT_CHAN_CENTER; break;
200 case 2: mode = MODE_2; channel_config = AOUT_CHANS_STEREO; break;
201 case 3: mode = MODE_1_2; channel_config = AOUT_CHANS_3_0; break;
202 case 4: mode = MODE_1_2_1; channel_config = AOUT_CHANS_4_CENTER_REAR; break;
203 case 5: mode = MODE_1_2_2; channel_config = AOUT_CHANS_5_0; break;
204 case 6: mode = MODE_1_2_2_1; channel_config = AOUT_CHANS_5_1; break;
205 case 8: mode = MODE_1_2_2_2_1; channel_config = AOUT_CHANS_7_1; break;
207 msg_Err(p_enc, "we do not support > 8 input channels, this input has %i",
208 p_enc->fmt_in.audio.i_channels);
212 p_enc->fmt_in.audio.i_physical_channels = channel_config;
214 msg_Info(p_enc, "Initializing AAC Encoder, %i channels", p_enc->fmt_in.audio.i_channels);
216 /* Allocate the memory needed to store the encoder's structure */
217 encoder_sys_t *p_sys = (encoder_sys_t *)malloc(sizeof(encoder_sys_t));
218 if (unlikely(!p_sys))
220 p_enc->p_sys = p_sys;
221 p_enc->fmt_in.i_codec = VLC_CODEC_S16N;
222 p_enc->fmt_out.i_cat = AUDIO_ES;
223 p_enc->fmt_out.i_codec = VLC_CODEC_MP4A;
225 p_sys->i_pts_last = 0;
228 erraac = aacEncOpen(&p_sys->handle, 0, p_enc->fmt_in.audio.i_channels);
229 if (erraac != AACENC_OK) {
230 msg_Err(p_enc, "Unable to open encoder: %s", fdkaac_error(erraac));
235 #define SET_PARAM(P, V) do { \
236 AACENC_ERROR err = aacEncoder_SetParam(p_sys->handle, AACENC_ ## P, V); \
237 if (err != AACENC_OK) { \
238 msg_Err(p_enc, "Couldn't set " #P " to value %d: %s", V, fdkaac_error(err)); \
243 SET_PARAM(AOT, i_aot);
244 bool b_eld_sbr = var_InheritBool(p_enc, ENC_CFG_PREFIX "sbr");
245 if (i_aot == PROFILE_AAC_ELD && b_eld_sbr)
246 SET_PARAM(SBR_MODE, 1);
247 SET_PARAM(SAMPLERATE, p_enc->fmt_out.audio.i_rate);
248 SET_PARAM(CHANNELMODE, mode);
249 SET_PARAM(CHANNELORDER, CH_ORDER_WG4);
251 int i_vbr = var_InheritInteger(p_enc, ENC_CFG_PREFIX "vbr");
253 if ((i_aot == PROFILE_AAC_HE || i_aot == PROFILE_AAC_HE_v2) && i_vbr > 3) {
254 msg_Warn(p_enc, "Maximum VBR quality for this profile is 3, setting vbr=3");
257 SET_PARAM(BITRATEMODE, i_vbr);
259 int i_bitrate = p_enc->fmt_out.i_bitrate;
260 if (i_bitrate == 0) {
261 i_bitrate = 96 * p_enc->fmt_in.audio.i_channels * p_enc->fmt_out.audio.i_rate / 44;
262 if (i_aot == PROFILE_AAC_HE || i_aot == PROFILE_AAC_HE_v2 || b_eld_sbr)
264 p_enc->fmt_out.i_bitrate = i_bitrate;
265 msg_Info(p_enc, "Setting optimal bitrate of %i", i_bitrate);
267 SET_PARAM(BITRATE, i_bitrate);
269 SET_PARAM(TRANSMUX, 0);
270 SET_PARAM(SIGNALING_MODE, (int)var_InheritInteger(p_enc, ENC_CFG_PREFIX "signaling"));
271 SET_PARAM(AFTERBURNER, !!var_InheritBool(p_enc, ENC_CFG_PREFIX "afterburner"));
274 erraac = aacEncEncode(p_sys->handle, NULL, NULL, NULL, NULL);
275 if (erraac != AACENC_OK) {
276 msg_Err(p_enc, "Unable to initialize the encoder: %s", fdkaac_error(erraac));
280 AACENC_InfoStruct info = { 0 };
281 erraac = aacEncInfo(p_sys->handle, &info);
282 if (erraac != AACENC_OK) {
283 msg_Err(p_enc, "Unable to get the encoder info: %s", fdkaac_error(erraac));
287 /* The maximum packet size is 6144 bits aka 768 bytes per channel. */
288 p_sys->i_maxoutputsize = 768*p_enc->fmt_in.audio.i_channels;
289 p_enc->fmt_in.audio.i_bitspersample = 16;
290 p_sys->i_frame_size = info.frameLength;
291 p_sys->i_encoderdelay = info.encoderDelay;
293 p_enc->fmt_out.i_extra = info.confSize;
294 if (p_enc->fmt_out.i_extra) {
295 p_enc->fmt_out.p_extra = malloc(p_enc->fmt_out.i_extra);
296 if (p_enc->fmt_out.p_extra == NULL) {
297 msg_Err(p_enc, "Unable to allocate fmt_out.p_extra");
300 memcpy(p_enc->fmt_out.p_extra, info.confBuf, p_enc->fmt_out.i_extra);
303 p_enc->pf_encode_audio = EncodeAudio;
306 // TODO: Add more debug info to this config printout
307 msg_Dbg(p_enc, "fmt_out.p_extra = %i", p_enc->fmt_out.i_extra);
313 CloseEncoder(p_this);
317 /****************************************************************************
318 * EncodeAudio: the whole thing
319 ****************************************************************************/
320 static block_t *EncodeAudio(encoder_t *p_enc, block_t *p_aout_buf)
326 encoder_sys_t *p_sys = p_enc->p_sys;
328 if (likely(p_aout_buf)) {
329 p_buffer = (int16_t *)p_aout_buf->p_buffer;
330 i_samples = p_aout_buf->i_nb_samples;
331 i_pts_out = p_aout_buf->i_pts - (mtime_t)((double)CLOCK_FREQ *
332 (double)p_sys->i_encoderdelay /
333 (double)p_enc->fmt_out.audio.i_rate);
334 if (p_sys->i_pts_last == 0)
335 p_sys->i_pts_last = i_pts_out - (mtime_t)((double)CLOCK_FREQ *
336 (double)(p_sys->i_frame_size) /
337 (double)p_enc->fmt_out.audio.i_rate);
340 i_pts_out = p_sys->i_pts_last;
343 int i_samples_left = i_samples;
344 int i_loop_count = 0;
346 block_t *p_chain = NULL;
347 while (i_samples_left >= 0) {
348 AACENC_BufDesc in_buf = { 0 }, out_buf = { 0 };
349 AACENC_InArgs in_args = { 0 };
350 AACENC_OutArgs out_args = { 0 };
351 int in_identifier = IN_AUDIO_DATA;
352 int in_size, in_elem_size;
353 int out_identifier = OUT_BITSTREAM_DATA;
354 int out_size, out_elem_size;
355 void *in_ptr, *out_ptr;
357 if (unlikely(i_samples == 0)) {
358 // this forces the encoder to purge whatever is left in the internal buffer
359 in_args.numInSamples = -1;
361 in_ptr = p_buffer + (i_samples - i_samples_left)*p_enc->fmt_in.audio.i_channels;
362 in_size = 2*p_enc->fmt_in.audio.i_channels*i_samples_left;
364 in_args.numInSamples = p_enc->fmt_in.audio.i_channels*i_samples_left;
366 in_buf.bufs = &in_ptr;
367 in_buf.bufferIdentifiers = &in_identifier;
368 in_buf.bufSizes = &in_size;
369 in_buf.bufElSizes = &in_elem_size;
372 p_block = block_Alloc(p_sys->i_maxoutputsize);
373 p_block->i_buffer = p_sys->i_maxoutputsize;
374 out_ptr = p_block->p_buffer;
375 out_size = p_block->i_buffer;
378 out_buf.bufs = &out_ptr;
379 out_buf.bufferIdentifiers = &out_identifier;
380 out_buf.bufSizes = &out_size;
381 out_buf.bufElSizes = &out_elem_size;
384 if ((erraac = aacEncEncode(p_sys->handle, &in_buf, &out_buf, &in_args, &out_args)) != AACENC_OK) {
385 if (erraac == AACENC_ENCODE_EOF) {
386 msg_Info(p_enc, "Encoding final bytes (EOF)");
388 msg_Err(p_enc, "Encoding failed: %s", fdkaac_error(erraac));
389 block_Release(p_block);
393 if (out_args.numOutBytes > 0) {
394 p_block->i_buffer = out_args.numOutBytes;
395 if (unlikely(i_samples == 0)) {
396 // I only have the numOutBytes so approximate based on compression factor
397 double d_samples_forward = p_sys->d_compression_ratio*(double)out_args.numOutBytes;
398 i_pts_out += (mtime_t)d_samples_forward;
399 p_block->i_length = (mtime_t)d_samples_forward;
400 // TODO: It would be more precise (a few microseconds) to use d_samples_forward =
401 // (mtime_t)CLOCK_FREQ * (mtime_t)p_sys->i_frame_size/(mtime_t)p_enc->fmt_out.audio.i_rate
402 // but I am not sure if the lib always outputs a full frame when
403 // emptying the internal buffer in the EOF scenario
405 if (i_loop_count == 0) {
406 // There can be an implicit delay in the first loop cycle because leftover bytes
407 // in the library buffer from the prior block
408 double d_samples_delay = (double)p_sys->i_frame_size - (double)out_args.numInSamples /
409 (double)p_enc->fmt_in.audio.i_channels;
410 i_pts_out -= (mtime_t)((double)CLOCK_FREQ * d_samples_delay /
411 (double)p_enc->fmt_out.audio.i_rate);
412 p_block->i_length = (mtime_t)((double)CLOCK_FREQ * (double)p_sys->i_frame_size /
413 (double)p_enc->fmt_out.audio.i_rate);
414 p_block->i_nb_samples = d_samples_delay;
415 //p_block->i_length = i_pts_out - p_sys->i_pts_last;
417 double d_samples_forward = (double)out_args.numInSamples/(double)p_enc->fmt_in.audio.i_channels;
418 double d_length = ((double)CLOCK_FREQ * d_samples_forward /
419 (double)p_enc->fmt_out.audio.i_rate);
420 i_pts_out += (mtime_t) d_length;
421 p_block->i_length = (mtime_t) d_length;
422 p_block->i_nb_samples = d_samples_forward;
425 p_block->i_dts = p_block->i_pts = i_pts_out;
426 block_ChainAppend(&p_chain, p_block);
428 msg_Dbg(p_enc, "dts %"PRId64", length %"PRId64", " "pts_last "
429 "%"PRId64" numOutBytes = %i, numInSamples = %i, "
430 "i_samples %i, i_loop_count %i",
431 p_block->i_dts, p_block->i_length,
432 p_sys->i_pts_last, out_args.numOutBytes,
433 out_args.numInSamples, i_samples, i_loop_count);
435 if (likely(i_samples > 0)) {
436 p_sys->d_compression_ratio = (double)p_block->i_length / (double)out_args.numOutBytes;
437 i_samples_left -= out_args.numInSamples/p_enc->fmt_in.audio.i_channels;
438 p_sys->i_pts_last = i_pts_out;
441 block_Release(p_block);
442 //msg_Dbg(p_enc, "aac_encode_audio: not enough data yet");
445 if (unlikely(i_loop_count++ > 100)) {
446 msg_Err(p_enc, "Loop count greater than 100!!!, something must be wrong with the encoder library");
454 /*****************************************************************************
455 * CloseDecoder: decoder destruction
456 *****************************************************************************/
457 static void CloseEncoder(vlc_object_t *p_this)
459 encoder_t *p_enc = (encoder_t *)p_this;
460 encoder_sys_t *p_sys = p_enc->p_sys;
462 aacEncClose(&p_sys->handle);