build: Fix Ogg demuxer dependencies
[libav.git] / libavcodec / libfdk-aacenc.c
CommitLineData
37eeb5e2
MS
1/*
2 * AAC encoder wrapper
3 * Copyright (c) 2012 Martin Storsjo
4 *
5 * This file is part of Libav.
6 *
7 * Libav 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.
11 *
12 * Libav 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.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with Libav; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22#include <fdk-aac/aacenc_lib.h>
23
24#include "avcodec.h"
25#include "audio_frame_queue.h"
26#include "internal.h"
27#include "libavutil/audioconvert.h"
28#include "libavutil/opt.h"
29
30typedef struct AACContext {
31 const AVClass *class;
32 HANDLE_AACENCODER handle;
33 int afterburner;
34 int eld_sbr;
35 int signaling;
36
37 AudioFrameQueue afq;
38} AACContext;
39
40static 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" },
48 { NULL }
49};
50
51static const AVClass aac_enc_class = {
52 "libfdk_aac", av_default_item_name, aac_enc_options, LIBAVUTIL_VERSION_INT
53};
54
55static const char *aac_get_error(AACENC_ERROR err)
56{
57 switch (err) {
58 case AACENC_OK:
59 return "No error";
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:
81 return "End of file";
82 default:
83 return "Unknown error";
84 }
85}
86
87static int aac_encode_close(AVCodecContext *avctx)
88{
89 AACContext *s = avctx->priv_data;
90
91 if (s->handle)
92 aacEncClose(&s->handle);
93#if FF_API_OLD_ENCODE_AUDIO
94 av_freep(&avctx->coded_frame);
95#endif
96 av_freep(&avctx->extradata);
97 ff_af_queue_close(&s->afq);
98
99 return 0;
100}
101
102static av_cold int aac_encode_init(AVCodecContext *avctx)
103{
104 AACContext *s = avctx->priv_data;
105 int ret = AVERROR(EINVAL);
106 AACENC_InfoStruct info = { 0 };
107 CHANNEL_MODE mode;
108 AACENC_ERROR err;
109 int aot = FF_PROFILE_AAC_LOW + 1;
110 int sce = 0, cpe = 0;
111
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",
114 aac_get_error(err));
115 goto error;
116 }
117
118 if (avctx->profile != FF_PROFILE_UNKNOWN)
119 aot = avctx->profile + 1;
120
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));
124 goto error;
125 }
126
127 if (aot == FF_PROFILE_AAC_ELD + 1 && s->eld_sbr) {
128 if ((err = aacEncoder_SetParam(s->handle, AACENC_SBR_MODE,
129 1)) != AACENC_OK) {
130 av_log(avctx, AV_LOG_ERROR, "Unable to enable SBR for ELD: %s\n",
131 aac_get_error(err));
132 goto error;
133 }
134 }
135
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));
140 goto error;
141 }
142
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;
150 default:
151 av_log(avctx, AV_LOG_ERROR,
152 "Unsupported number of channels %d\n", avctx->channels);
153 goto error;
154 }
155
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));
160 goto error;
161 }
162
163 if ((err = aacEncoder_SetParam(s->handle, AACENC_CHANNELORDER,
164 1)) != AACENC_OK) {
165 av_log(avctx, AV_LOG_ERROR,
166 "Unable to set wav channel order %d: %s\n",
167 mode, aac_get_error(err));
168 goto error;
169 }
170
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);
177 }
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));
182 goto error;
183 }
184 } else {
185 if (avctx->bit_rate <= 0) {
186 if (avctx->profile == FF_PROFILE_AAC_HE_V2) {
187 sce = 1;
188 cpe = 0;
189 }
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 ||
193 s->eld_sbr)
194 avctx->bit_rate /= 2;
195 }
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));
200 goto error;
201 }
202 }
203
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",
209 aac_get_error(err));
210 goto error;
211 }
212
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;
218
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));
223 goto error;
224 }
225
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));
230 goto error;
231 }
232
233 if ((err = aacEncEncode(s->handle, NULL, NULL, NULL, NULL)) != AACENC_OK) {
234 av_log(avctx, AV_LOG_ERROR, "Unable to initialize the encoder: %s\n",
235 aac_get_error(err));
236 return AVERROR(EINVAL);
237 }
238
239 if ((err = aacEncInfo(s->handle, &info)) != AACENC_OK) {
240 av_log(avctx, AV_LOG_ERROR, "Unable to get encoder info: %s\n",
241 aac_get_error(err));
242 goto error;
243 }
244
245#if FF_API_OLD_ENCODE_AUDIO
246 avctx->coded_frame = avcodec_alloc_frame();
247 if (!avctx->coded_frame) {
248 ret = AVERROR(ENOMEM);
249 goto error;
250 }
251#endif
252 avctx->frame_size = info.frameLength;
253 avctx->delay = info.encoderDelay;
254 ff_af_queue_init(avctx, &s->afq);
255
256 if (avctx->flags & CODEC_FLAG_GLOBAL_HEADER) {
257 avctx->extradata_size = info.confSize;
258 avctx->extradata = av_mallocz(avctx->extradata_size +
259 FF_INPUT_BUFFER_PADDING_SIZE);
260 if (!avctx->extradata) {
261 ret = AVERROR(ENOMEM);
262 goto error;
263 }
264
265 memcpy(avctx->extradata, info.confBuf, info.confSize);
266 }
267 return 0;
268error:
269 aac_encode_close(avctx);
270 return ret;
271}
272
273static int aac_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
274 const AVFrame *frame, int *got_packet_ptr)
275{
276 AACContext *s = avctx->priv_data;
277 AACENC_BufDesc in_buf = { 0 }, out_buf = { 0 };
278 AACENC_InArgs in_args = { 0 };
279 AACENC_OutArgs out_args = { 0 };
280 int in_buffer_identifier = IN_AUDIO_DATA;
281 int in_buffer_size, in_buffer_element_size;
282 int out_buffer_identifier = OUT_BITSTREAM_DATA;
283 int out_buffer_size, out_buffer_element_size;
284 void *in_ptr, *out_ptr;
285 int ret;
286 AACENC_ERROR err;
287
288 /* handle end-of-stream small frame and flushing */
289 if (!frame) {
290 in_args.numInSamples = -1;
291 } else {
292 in_ptr = frame->data[0];
293 in_buffer_size = 2 * avctx->channels * frame->nb_samples;
294 in_buffer_element_size = 2;
295
296 in_args.numInSamples = avctx->channels * frame->nb_samples;
297 in_buf.numBufs = 1;
298 in_buf.bufs = &in_ptr;
299 in_buf.bufferIdentifiers = &in_buffer_identifier;
300 in_buf.bufSizes = &in_buffer_size;
301 in_buf.bufElSizes = &in_buffer_element_size;
302
303 /* add current frame to the queue */
304 if ((ret = ff_af_queue_add(&s->afq, frame) < 0))
305 return ret;
306 }
307
308 /* The maximum packet size is 6144 bits aka 768 bytes per channel. */
309 if ((ret = ff_alloc_packet(avpkt, FFMAX(8192, 768 * avctx->channels)))) {
310 av_log(avctx, AV_LOG_ERROR, "Error getting output packet\n");
311 return ret;
312 }
313
314 out_ptr = avpkt->data;
315 out_buffer_size = avpkt->size;
316 out_buffer_element_size = 1;
317 out_buf.numBufs = 1;
318 out_buf.bufs = &out_ptr;
319 out_buf.bufferIdentifiers = &out_buffer_identifier;
320 out_buf.bufSizes = &out_buffer_size;
321 out_buf.bufElSizes = &out_buffer_element_size;
322
323 if ((err = aacEncEncode(s->handle, &in_buf, &out_buf, &in_args,
324 &out_args)) != AACENC_OK) {
325 if (!frame && err == AACENC_ENCODE_EOF)
326 return 0;
327 av_log(avctx, AV_LOG_ERROR, "Unable to encode frame: %s\n",
328 aac_get_error(err));
329 return AVERROR(EINVAL);
330 }
331
332 if (!out_args.numOutBytes)
333 return 0;
334
335 /* Get the next frame pts & duration */
336 ff_af_queue_remove(&s->afq, avctx->frame_size, &avpkt->pts,
337 &avpkt->duration);
338
339 avpkt->size = out_args.numOutBytes;
340 *got_packet_ptr = 1;
341 return 0;
342}
343
344static const AVProfile profiles[] = {
345 { FF_PROFILE_AAC_LOW, "LC" },
346 { FF_PROFILE_AAC_HE, "HE-AAC" },
347 { FF_PROFILE_AAC_HE_V2, "HE-AACv2" },
348 { FF_PROFILE_AAC_LD, "LD" },
349 { FF_PROFILE_AAC_ELD, "ELD" },
350 { FF_PROFILE_UNKNOWN },
351};
352
353static const AVCodecDefault aac_encode_defaults[] = {
354 { "b", "0" },
355 { NULL }
356};
357
358static const uint64_t aac_channel_layout[] = {
359 AV_CH_LAYOUT_MONO,
360 AV_CH_LAYOUT_STEREO,
361 AV_CH_LAYOUT_SURROUND,
362 AV_CH_LAYOUT_4POINT0,
363 AV_CH_LAYOUT_5POINT0_BACK,
364 AV_CH_LAYOUT_5POINT1_BACK,
365 0,
366};
367
368AVCodec ff_libfdk_aac_encoder = {
369 .name = "libfdk_aac",
370 .type = AVMEDIA_TYPE_AUDIO,
371 .id = CODEC_ID_AAC,
372 .priv_data_size = sizeof(AACContext),
373 .init = aac_encode_init,
374 .encode2 = aac_encode_frame,
375 .close = aac_encode_close,
376 .capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY,
377 .sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
378 AV_SAMPLE_FMT_NONE },
379 .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
380 .priv_class = &aac_enc_class,
381 .defaults = aac_encode_defaults,
382 .profiles = profiles,
383 .channel_layouts = aac_channel_layout,
384};