3 * Copyright (c) 2012 Martin Storsjo
5 * This file is part of FFmpeg.
7 * FFmpeg 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.
12 * FFmpeg 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.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 #include <fdk-aac/aacenc_lib.h>
24 #include "libavutil/channel_layout.h"
25 #include "libavutil/common.h"
26 #include "libavutil/opt.h"
28 #include "audio_frame_queue.h"
31 typedef struct AACContext {
33 HANDLE_AACENCODER handle;
44 static const AVOption aac_enc_options[] = {
45 { "afterburner", "Afterburner (improved quality)", offsetof(AACContext, afterburner), AV_OPT_TYPE_INT, { .i64 = 1 }, 0, 1, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
46 { "eld_sbr", "Enable SBR for ELD (for SBR in other configurations, use the -profile parameter)", offsetof(AACContext, eld_sbr), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
47 { "signaling", "SBR/PS signaling style", offsetof(AACContext, signaling), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 2, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
48 { "default", "Choose signaling implicitly (explicit hierarchical by default, implicit if global header is disabled)", 0, AV_OPT_TYPE_CONST, { .i64 = -1 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
49 { "implicit", "Implicit backwards compatible signaling", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
50 { "explicit_sbr", "Explicit SBR, implicit PS signaling", 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
51 { "explicit_hierarchical", "Explicit hierarchical signaling", 0, AV_OPT_TYPE_CONST, { .i64 = 2 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
52 { "latm", "Output LATM/LOAS encapsulated data", offsetof(AACContext, latm), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
53 { "header_period", "StreamMuxConfig and PCE repetition period (in frames)", offsetof(AACContext, header_period), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 0xffff, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
54 { "vbr", "VBR mode (1-5)", offsetof(AACContext, vbr), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 5, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
58 static const AVClass aac_enc_class = {
59 "libfdk_aac", av_default_item_name, aac_enc_options, LIBAVUTIL_VERSION_INT
62 static const char *aac_get_error(AACENC_ERROR err)
67 case AACENC_INVALID_HANDLE:
68 return "Invalid handle";
69 case AACENC_MEMORY_ERROR:
70 return "Memory allocation error";
71 case AACENC_UNSUPPORTED_PARAMETER:
72 return "Unsupported parameter";
73 case AACENC_INVALID_CONFIG:
74 return "Invalid config";
75 case AACENC_INIT_ERROR:
76 return "Initialization error";
77 case AACENC_INIT_AAC_ERROR:
78 return "AAC library initialization error";
79 case AACENC_INIT_SBR_ERROR:
80 return "SBR library initialization error";
81 case AACENC_INIT_TP_ERROR:
82 return "Transport library initialization error";
83 case AACENC_INIT_META_ERROR:
84 return "Metadata library initialization error";
85 case AACENC_ENCODE_ERROR:
86 return "Encoding error";
87 case AACENC_ENCODE_EOF:
90 return "Unknown error";
94 static int aac_encode_close(AVCodecContext *avctx)
96 AACContext *s = avctx->priv_data;
99 aacEncClose(&s->handle);
100 av_freep(&avctx->extradata);
101 ff_af_queue_close(&s->afq);
106 static av_cold int aac_encode_init(AVCodecContext *avctx)
108 AACContext *s = avctx->priv_data;
109 int ret = AVERROR(EINVAL);
110 AACENC_InfoStruct info = { 0 };
113 int aot = FF_PROFILE_AAC_LOW + 1;
114 int sce = 0, cpe = 0;
116 if ((err = aacEncOpen(&s->handle, 0, avctx->channels)) != AACENC_OK) {
117 av_log(avctx, AV_LOG_ERROR, "Unable to open the encoder: %s\n",
122 if (avctx->profile != FF_PROFILE_UNKNOWN)
123 aot = avctx->profile + 1;
125 if ((err = aacEncoder_SetParam(s->handle, AACENC_AOT, aot)) != AACENC_OK) {
126 av_log(avctx, AV_LOG_ERROR, "Unable to set the AOT %d: %s\n",
127 aot, aac_get_error(err));
131 if (aot == FF_PROFILE_AAC_ELD + 1 && s->eld_sbr) {
132 if ((err = aacEncoder_SetParam(s->handle, AACENC_SBR_MODE,
134 av_log(avctx, AV_LOG_ERROR, "Unable to enable SBR for ELD: %s\n",
140 if ((err = aacEncoder_SetParam(s->handle, AACENC_SAMPLERATE,
141 avctx->sample_rate)) != AACENC_OK) {
142 av_log(avctx, AV_LOG_ERROR, "Unable to set the sample rate %d: %s\n",
143 avctx->sample_rate, aac_get_error(err));
147 switch (avctx->channels) {
148 case 1: mode = MODE_1; sce = 1; cpe = 0; break;
149 case 2: mode = MODE_2; sce = 0; cpe = 1; break;
150 case 3: mode = MODE_1_2; sce = 1; cpe = 1; break;
151 case 4: mode = MODE_1_2_1; sce = 2; cpe = 1; break;
152 case 5: mode = MODE_1_2_2; sce = 1; cpe = 2; break;
153 case 6: mode = MODE_1_2_2_1; sce = 2; cpe = 2; break;
154 /* The version macro is introduced the same time as the 7.1 support, so this
156 #ifdef AACENCODER_LIB_VL0
160 if (avctx->channel_layout == AV_CH_LAYOUT_7POINT1) {
161 mode = MODE_7_1_REAR_SURROUND;
163 // MODE_1_2_2_2_1 and MODE_7_1_FRONT_CENTER use the same channel layout
164 mode = MODE_7_1_FRONT_CENTER;
169 av_log(avctx, AV_LOG_ERROR,
170 "Unsupported number of channels %d\n", avctx->channels);
174 if ((err = aacEncoder_SetParam(s->handle, AACENC_CHANNELMODE,
175 mode)) != AACENC_OK) {
176 av_log(avctx, AV_LOG_ERROR,
177 "Unable to set channel mode %d: %s\n", mode, aac_get_error(err));
181 if ((err = aacEncoder_SetParam(s->handle, AACENC_CHANNELORDER,
183 av_log(avctx, AV_LOG_ERROR,
184 "Unable to set wav channel order %d: %s\n",
185 mode, aac_get_error(err));
189 if (avctx->flags & CODEC_FLAG_QSCALE || s->vbr) {
190 int mode = s->vbr ? s->vbr : avctx->global_quality;
191 if (mode < 1 || mode > 5) {
192 av_log(avctx, AV_LOG_WARNING,
193 "VBR quality %d out of range, should be 1-5\n", mode);
194 mode = av_clip(mode, 1, 5);
196 av_log(avctx, AV_LOG_WARNING,
197 "Note, the VBR setting is unsupported and only works with "
198 "some parameter combinations\n");
199 if ((err = aacEncoder_SetParam(s->handle, AACENC_BITRATEMODE,
200 mode)) != AACENC_OK) {
201 av_log(avctx, AV_LOG_ERROR, "Unable to set the VBR bitrate mode %d: %s\n",
202 mode, aac_get_error(err));
206 if (avctx->bit_rate <= 0) {
207 if (avctx->profile == FF_PROFILE_AAC_HE_V2) {
211 avctx->bit_rate = (96*sce + 128*cpe) * avctx->sample_rate / 44;
212 if (avctx->profile == FF_PROFILE_AAC_HE ||
213 avctx->profile == FF_PROFILE_AAC_HE_V2 ||
214 avctx->profile == FF_PROFILE_MPEG2_AAC_HE ||
216 avctx->bit_rate /= 2;
218 if ((err = aacEncoder_SetParam(s->handle, AACENC_BITRATE,
219 avctx->bit_rate)) != AACENC_OK) {
220 av_log(avctx, AV_LOG_ERROR, "Unable to set the bitrate %d: %s\n",
221 avctx->bit_rate, aac_get_error(err));
226 /* Choose bitstream format - if global header is requested, use
227 * raw access units, otherwise use ADTS. */
228 if ((err = aacEncoder_SetParam(s->handle, AACENC_TRANSMUX,
229 avctx->flags & CODEC_FLAG_GLOBAL_HEADER ? 0 : s->latm ? 10 : 2)) != AACENC_OK) {
230 av_log(avctx, AV_LOG_ERROR, "Unable to set the transmux format: %s\n",
235 if (s->latm && s->header_period) {
236 if ((err = aacEncoder_SetParam(s->handle, AACENC_HEADER_PERIOD,
237 s->header_period)) != AACENC_OK) {
238 av_log(avctx, AV_LOG_ERROR, "Unable to set header period: %s\n",
244 /* If no signaling mode is chosen, use explicit hierarchical signaling
245 * if using mp4 mode (raw access units, with global header) and
246 * implicit signaling if using ADTS. */
247 if (s->signaling < 0)
248 s->signaling = avctx->flags & CODEC_FLAG_GLOBAL_HEADER ? 2 : 0;
250 if ((err = aacEncoder_SetParam(s->handle, AACENC_SIGNALING_MODE,
251 s->signaling)) != AACENC_OK) {
252 av_log(avctx, AV_LOG_ERROR, "Unable to set signaling mode %d: %s\n",
253 s->signaling, aac_get_error(err));
257 if ((err = aacEncoder_SetParam(s->handle, AACENC_AFTERBURNER,
258 s->afterburner)) != AACENC_OK) {
259 av_log(avctx, AV_LOG_ERROR, "Unable to set afterburner to %d: %s\n",
260 s->afterburner, aac_get_error(err));
264 if (avctx->cutoff > 0) {
265 if (avctx->cutoff < (avctx->sample_rate + 255) >> 8 || avctx->cutoff > 20000) {
266 av_log(avctx, AV_LOG_ERROR, "cutoff valid range is %d-20000\n",
267 (avctx->sample_rate + 255) >> 8);
270 if ((err = aacEncoder_SetParam(s->handle, AACENC_BANDWIDTH,
271 avctx->cutoff)) != AACENC_OK) {
272 av_log(avctx, AV_LOG_ERROR, "Unable to set the encoder bandwidth to %d: %s\n",
273 avctx->cutoff, aac_get_error(err));
278 if ((err = aacEncEncode(s->handle, NULL, NULL, NULL, NULL)) != AACENC_OK) {
279 av_log(avctx, AV_LOG_ERROR, "Unable to initialize the encoder: %s\n",
281 return AVERROR(EINVAL);
284 if ((err = aacEncInfo(s->handle, &info)) != AACENC_OK) {
285 av_log(avctx, AV_LOG_ERROR, "Unable to get encoder info: %s\n",
290 avctx->frame_size = info.frameLength;
291 avctx->delay = info.encoderDelay;
292 ff_af_queue_init(avctx, &s->afq);
294 if (avctx->flags & CODEC_FLAG_GLOBAL_HEADER) {
295 avctx->extradata_size = info.confSize;
296 avctx->extradata = av_mallocz(avctx->extradata_size +
297 FF_INPUT_BUFFER_PADDING_SIZE);
298 if (!avctx->extradata) {
299 ret = AVERROR(ENOMEM);
303 memcpy(avctx->extradata, info.confBuf, info.confSize);
307 aac_encode_close(avctx);
311 static int aac_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
312 const AVFrame *frame, int *got_packet_ptr)
314 AACContext *s = avctx->priv_data;
315 AACENC_BufDesc in_buf = { 0 }, out_buf = { 0 };
316 AACENC_InArgs in_args = { 0 };
317 AACENC_OutArgs out_args = { 0 };
318 int in_buffer_identifier = IN_AUDIO_DATA;
319 int in_buffer_size, in_buffer_element_size;
320 int out_buffer_identifier = OUT_BITSTREAM_DATA;
321 int out_buffer_size, out_buffer_element_size;
322 void *in_ptr, *out_ptr;
326 /* handle end-of-stream small frame and flushing */
328 in_args.numInSamples = -1;
330 in_ptr = frame->data[0];
331 in_buffer_size = 2 * avctx->channels * frame->nb_samples;
332 in_buffer_element_size = 2;
334 in_args.numInSamples = avctx->channels * frame->nb_samples;
336 in_buf.bufs = &in_ptr;
337 in_buf.bufferIdentifiers = &in_buffer_identifier;
338 in_buf.bufSizes = &in_buffer_size;
339 in_buf.bufElSizes = &in_buffer_element_size;
341 /* add current frame to the queue */
342 if ((ret = ff_af_queue_add(&s->afq, frame)) < 0)
346 /* The maximum packet size is 6144 bits aka 768 bytes per channel. */
347 if ((ret = ff_alloc_packet2(avctx, avpkt, FFMAX(8192, 768 * avctx->channels))) < 0)
350 out_ptr = avpkt->data;
351 out_buffer_size = avpkt->size;
352 out_buffer_element_size = 1;
354 out_buf.bufs = &out_ptr;
355 out_buf.bufferIdentifiers = &out_buffer_identifier;
356 out_buf.bufSizes = &out_buffer_size;
357 out_buf.bufElSizes = &out_buffer_element_size;
359 if ((err = aacEncEncode(s->handle, &in_buf, &out_buf, &in_args,
360 &out_args)) != AACENC_OK) {
361 if (!frame && err == AACENC_ENCODE_EOF)
363 av_log(avctx, AV_LOG_ERROR, "Unable to encode frame: %s\n",
365 return AVERROR(EINVAL);
368 if (!out_args.numOutBytes)
371 /* Get the next frame pts & duration */
372 ff_af_queue_remove(&s->afq, avctx->frame_size, &avpkt->pts,
375 avpkt->size = out_args.numOutBytes;
380 static const AVProfile profiles[] = {
381 { FF_PROFILE_AAC_LOW, "LC" },
382 { FF_PROFILE_AAC_HE, "HE-AAC" },
383 { FF_PROFILE_AAC_HE_V2, "HE-AACv2" },
384 { FF_PROFILE_AAC_LD, "LD" },
385 { FF_PROFILE_AAC_ELD, "ELD" },
386 { FF_PROFILE_UNKNOWN },
389 static const AVCodecDefault aac_encode_defaults[] = {
394 static const uint64_t aac_channel_layout[] = {
397 AV_CH_LAYOUT_SURROUND,
398 AV_CH_LAYOUT_4POINT0,
399 AV_CH_LAYOUT_5POINT0_BACK,
400 AV_CH_LAYOUT_5POINT1_BACK,
401 #ifdef AACENCODER_LIB_VL0
402 AV_CH_LAYOUT_7POINT1_WIDE_BACK,
403 AV_CH_LAYOUT_7POINT1,
408 static const int aac_sample_rates[] = {
409 96000, 88200, 64000, 48000, 44100, 32000,
410 24000, 22050, 16000, 12000, 11025, 8000, 0
413 AVCodec ff_libfdk_aac_encoder = {
414 .name = "libfdk_aac",
415 .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
416 .type = AVMEDIA_TYPE_AUDIO,
417 .id = AV_CODEC_ID_AAC,
418 .priv_data_size = sizeof(AACContext),
419 .init = aac_encode_init,
420 .encode2 = aac_encode_frame,
421 .close = aac_encode_close,
422 .capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY,
423 .sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
424 AV_SAMPLE_FMT_NONE },
425 .priv_class = &aac_enc_class,
426 .defaults = aac_encode_defaults,
427 .profiles = profiles,
428 .supported_samplerates = aac_sample_rates,
429 .channel_layouts = aac_channel_layout,