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>
25 #include "audio_frame_queue.h"
27 #include "libavutil/audioconvert.h"
28 #include "libavutil/opt.h"
30 typedef struct AACContext {
32 HANDLE_AACENCODER handle;
40 static const AVOption aac_enc_options[] = {
41 { "afterburner", "Afterburner (improved quality)", offsetof(AACContext, afterburner), AV_OPT_TYPE_INT, { 1 }, 0, 1, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
42 { "eld_sbr", "Enable SBR for ELD (for SBR in other configurations, use the -profile parameter)", offsetof(AACContext, eld_sbr), AV_OPT_TYPE_INT, { 0 }, 0, 1, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM },
43 { "signaling", "SBR/PS signaling style", offsetof(AACContext, signaling), AV_OPT_TYPE_INT, { -1 }, -1, 2, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
44 { "default", "Choose signaling implicitly (explicit hierarchical by default, implicit if global header is disabled)", 0, AV_OPT_TYPE_CONST, { -1 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
45 { "implicit", "Implicit backwards compatible signaling", 0, AV_OPT_TYPE_CONST, { 0 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
46 { "explicit_sbr", "Explicit SBR, implicit PS signaling", 0, AV_OPT_TYPE_CONST, { 1 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
47 { "explicit_hierarchical", "Explicit hierarchical signaling", 0, AV_OPT_TYPE_CONST, { 2 }, 0, 0, AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM, "signaling" },
51 static const AVClass aac_enc_class = {
52 "libfdk_aac", av_default_item_name, aac_enc_options, LIBAVUTIL_VERSION_INT
55 static const char *aac_get_error(AACENC_ERROR err)
60 case AACENC_INVALID_HANDLE:
61 return "Invalid handle";
62 case AACENC_MEMORY_ERROR:
63 return "Memory allocation error";
64 case AACENC_UNSUPPORTED_PARAMETER:
65 return "Unsupported parameter";
66 case AACENC_INVALID_CONFIG:
67 return "Invalid config";
68 case AACENC_INIT_ERROR:
69 return "Initialization error";
70 case AACENC_INIT_AAC_ERROR:
71 return "AAC library initialization error";
72 case AACENC_INIT_SBR_ERROR:
73 return "SBR library initialization error";
74 case AACENC_INIT_TP_ERROR:
75 return "Transport library initialization error";
76 case AACENC_INIT_META_ERROR:
77 return "Metadata library initialization error";
78 case AACENC_ENCODE_ERROR:
79 return "Encoding error";
80 case AACENC_ENCODE_EOF:
83 return "Unknown error";
87 static int aac_encode_close(AVCodecContext *avctx)
89 AACContext *s = avctx->priv_data;
92 aacEncClose(&s->handle);
93 #if FF_API_OLD_ENCODE_AUDIO
94 av_freep(&avctx->coded_frame);
96 av_freep(&avctx->extradata);
97 ff_af_queue_close(&s->afq);
102 static av_cold int aac_encode_init(AVCodecContext *avctx)
104 AACContext *s = avctx->priv_data;
105 int ret = AVERROR(EINVAL);
106 AACENC_InfoStruct info = { 0 };
109 int aot = FF_PROFILE_AAC_LOW + 1;
110 int sce = 0, cpe = 0;
112 if ((err = aacEncOpen(&s->handle, 0, avctx->channels)) != AACENC_OK) {
113 av_log(avctx, AV_LOG_ERROR, "Unable to open the encoder: %s\n",
118 if (avctx->profile != FF_PROFILE_UNKNOWN)
119 aot = avctx->profile + 1;
121 if ((err = aacEncoder_SetParam(s->handle, AACENC_AOT, aot)) != AACENC_OK) {
122 av_log(avctx, AV_LOG_ERROR, "Unable to set the AOT %d: %s\n",
123 aot, aac_get_error(err));
127 if (aot == FF_PROFILE_AAC_ELD + 1 && s->eld_sbr) {
128 if ((err = aacEncoder_SetParam(s->handle, AACENC_SBR_MODE,
130 av_log(avctx, AV_LOG_ERROR, "Unable to enable SBR for ELD: %s\n",
136 if ((err = aacEncoder_SetParam(s->handle, AACENC_SAMPLERATE,
137 avctx->sample_rate)) != AACENC_OK) {
138 av_log(avctx, AV_LOG_ERROR, "Unable to set the sample rate %d: %s\n",
139 avctx->sample_rate, aac_get_error(err));
143 switch (avctx->channels) {
144 case 1: mode = MODE_1; sce = 1; cpe = 0; break;
145 case 2: mode = MODE_2; sce = 0; cpe = 1; break;
146 case 3: mode = MODE_1_2; sce = 1; cpe = 1; break;
147 case 4: mode = MODE_1_2_1; sce = 2; cpe = 1; break;
148 case 5: mode = MODE_1_2_2; sce = 1; cpe = 2; break;
149 case 6: mode = MODE_1_2_2_1; sce = 2; cpe = 2; break;
151 av_log(avctx, AV_LOG_ERROR,
152 "Unsupported number of channels %d\n", avctx->channels);
156 if ((err = aacEncoder_SetParam(s->handle, AACENC_CHANNELMODE,
157 mode)) != AACENC_OK) {
158 av_log(avctx, AV_LOG_ERROR,
159 "Unable to set channel mode %d: %s\n", mode, aac_get_error(err));
163 if ((err = aacEncoder_SetParam(s->handle, AACENC_CHANNELORDER,
165 av_log(avctx, AV_LOG_ERROR,
166 "Unable to set wav channel order %d: %s\n",
167 mode, aac_get_error(err));
171 if (avctx->flags & CODEC_FLAG_QSCALE) {
172 int mode = avctx->global_quality;
173 if (mode < 1 || mode > 5) {
174 av_log(avctx, AV_LOG_WARNING,
175 "VBR quality %d out of range, should be 1-5\n", mode);
176 mode = av_clip(mode, 1, 5);
178 if ((err = aacEncoder_SetParam(s->handle, AACENC_BITRATEMODE,
179 mode)) != AACENC_OK) {
180 av_log(avctx, AV_LOG_ERROR, "Unable to set the VBR bitrate mode %d: %s\n",
181 mode, aac_get_error(err));
185 if (avctx->bit_rate <= 0) {
186 if (avctx->profile == FF_PROFILE_AAC_HE_V2) {
190 avctx->bit_rate = (96*sce + 128*cpe) * avctx->sample_rate / 44;
191 if (avctx->profile == FF_PROFILE_AAC_HE ||
192 avctx->profile == FF_PROFILE_AAC_HE_V2 ||
194 avctx->bit_rate /= 2;
196 if ((err = aacEncoder_SetParam(s->handle, AACENC_BITRATE,
197 avctx->bit_rate)) != AACENC_OK) {
198 av_log(avctx, AV_LOG_ERROR, "Unable to set the bitrate %d: %s\n",
199 avctx->bit_rate, aac_get_error(err));
204 /* Choose bitstream format - if global header is requested, use
205 * raw access units, otherwise use ADTS. */
206 if ((err = aacEncoder_SetParam(s->handle, AACENC_TRANSMUX,
207 avctx->flags & CODEC_FLAG_GLOBAL_HEADER ? 0 : 2)) != AACENC_OK) {
208 av_log(avctx, AV_LOG_ERROR, "Unable to set the transmux format: %s\n",
213 /* If no signaling mode is chosen, use explicit hierarchical signaling
214 * if using mp4 mode (raw access units, with global header) and
215 * implicit signaling if using ADTS. */
216 if (s->signaling < 0)
217 s->signaling = avctx->flags & CODEC_FLAG_GLOBAL_HEADER ? 2 : 0;
219 if ((err = aacEncoder_SetParam(s->handle, AACENC_SIGNALING_MODE,
220 s->signaling)) != AACENC_OK) {
221 av_log(avctx, AV_LOG_ERROR, "Unable to set signaling mode %d: %s\n",
222 s->signaling, aac_get_error(err));
226 if ((err = aacEncoder_SetParam(s->handle, AACENC_AFTERBURNER,
227 s->afterburner)) != AACENC_OK) {
228 av_log(avctx, AV_LOG_ERROR, "Unable to set afterburner to %d: %s\n",
229 s->afterburner, aac_get_error(err));
233 if (avctx->cutoff > 0) {
234 if ((err = aacEncoder_SetParam(s->handle, AACENC_BANDWIDTH,
235 avctx->cutoff)) != AACENC_OK) {
236 av_log(avctx, AV_LOG_ERROR, "Unable to set the encoder bandwith to %d: %s\n",
237 avctx->cutoff, aac_get_error(err));
242 if ((err = aacEncEncode(s->handle, NULL, NULL, NULL, NULL)) != AACENC_OK) {
243 av_log(avctx, AV_LOG_ERROR, "Unable to initialize the encoder: %s\n",
245 return AVERROR(EINVAL);
248 if ((err = aacEncInfo(s->handle, &info)) != AACENC_OK) {
249 av_log(avctx, AV_LOG_ERROR, "Unable to get encoder info: %s\n",
254 #if FF_API_OLD_ENCODE_AUDIO
255 avctx->coded_frame = avcodec_alloc_frame();
256 if (!avctx->coded_frame) {
257 ret = AVERROR(ENOMEM);
261 avctx->frame_size = info.frameLength;
262 avctx->delay = info.encoderDelay;
263 ff_af_queue_init(avctx, &s->afq);
265 if (avctx->flags & CODEC_FLAG_GLOBAL_HEADER) {
266 avctx->extradata_size = info.confSize;
267 avctx->extradata = av_mallocz(avctx->extradata_size +
268 FF_INPUT_BUFFER_PADDING_SIZE);
269 if (!avctx->extradata) {
270 ret = AVERROR(ENOMEM);
274 memcpy(avctx->extradata, info.confBuf, info.confSize);
278 aac_encode_close(avctx);
282 static int aac_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
283 const AVFrame *frame, int *got_packet_ptr)
285 AACContext *s = avctx->priv_data;
286 AACENC_BufDesc in_buf = { 0 }, out_buf = { 0 };
287 AACENC_InArgs in_args = { 0 };
288 AACENC_OutArgs out_args = { 0 };
289 int in_buffer_identifier = IN_AUDIO_DATA;
290 int in_buffer_size, in_buffer_element_size;
291 int out_buffer_identifier = OUT_BITSTREAM_DATA;
292 int out_buffer_size, out_buffer_element_size;
293 void *in_ptr, *out_ptr;
297 /* handle end-of-stream small frame and flushing */
299 in_args.numInSamples = -1;
301 in_ptr = frame->data[0];
302 in_buffer_size = 2 * avctx->channels * frame->nb_samples;
303 in_buffer_element_size = 2;
305 in_args.numInSamples = avctx->channels * frame->nb_samples;
307 in_buf.bufs = &in_ptr;
308 in_buf.bufferIdentifiers = &in_buffer_identifier;
309 in_buf.bufSizes = &in_buffer_size;
310 in_buf.bufElSizes = &in_buffer_element_size;
312 /* add current frame to the queue */
313 if ((ret = ff_af_queue_add(&s->afq, frame) < 0))
317 /* The maximum packet size is 6144 bits aka 768 bytes per channel. */
318 if ((ret = ff_alloc_packet(avpkt, FFMAX(8192, 768 * avctx->channels)))) {
319 av_log(avctx, AV_LOG_ERROR, "Error getting output packet\n");
323 out_ptr = avpkt->data;
324 out_buffer_size = avpkt->size;
325 out_buffer_element_size = 1;
327 out_buf.bufs = &out_ptr;
328 out_buf.bufferIdentifiers = &out_buffer_identifier;
329 out_buf.bufSizes = &out_buffer_size;
330 out_buf.bufElSizes = &out_buffer_element_size;
332 if ((err = aacEncEncode(s->handle, &in_buf, &out_buf, &in_args,
333 &out_args)) != AACENC_OK) {
334 if (!frame && err == AACENC_ENCODE_EOF)
336 av_log(avctx, AV_LOG_ERROR, "Unable to encode frame: %s\n",
338 return AVERROR(EINVAL);
341 if (!out_args.numOutBytes)
344 /* Get the next frame pts & duration */
345 ff_af_queue_remove(&s->afq, avctx->frame_size, &avpkt->pts,
348 avpkt->size = out_args.numOutBytes;
353 static const AVProfile profiles[] = {
354 { FF_PROFILE_AAC_LOW, "LC" },
355 { FF_PROFILE_AAC_HE, "HE-AAC" },
356 { FF_PROFILE_AAC_HE_V2, "HE-AACv2" },
357 { FF_PROFILE_AAC_LD, "LD" },
358 { FF_PROFILE_AAC_ELD, "ELD" },
359 { FF_PROFILE_UNKNOWN },
362 static const AVCodecDefault aac_encode_defaults[] = {
367 static const uint64_t aac_channel_layout[] = {
370 AV_CH_LAYOUT_SURROUND,
371 AV_CH_LAYOUT_4POINT0,
372 AV_CH_LAYOUT_5POINT0_BACK,
373 AV_CH_LAYOUT_5POINT1_BACK,
377 AVCodec ff_libfdk_aac_encoder = {
378 .name = "libfdk_aac",
379 .type = AVMEDIA_TYPE_AUDIO,
381 .priv_data_size = sizeof(AACContext),
382 .init = aac_encode_init,
383 .encode2 = aac_encode_frame,
384 .close = aac_encode_close,
385 .capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY,
386 .sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
387 AV_SAMPLE_FMT_NONE },
388 .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
389 .priv_class = &aac_enc_class,
390 .defaults = aac_encode_defaults,
391 .profiles = profiles,
392 .channel_layouts = aac_channel_layout,