1 /*
2 * AAC decoder wrapper
3 * Copyright (c) 2012 Martin Storsjo
4 *
5 * This file is part of FFmpeg.
6 *
7 * Permission to use, copy, modify, and/or distribute this software for any
8 * purpose with or without fee is hereby granted, provided that the above
9 * copyright notice and this permission notice appear in all copies.
10 *
11 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
12 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
13 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
14 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
15 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
17 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
18 */
19
20 #include <fdk-aac/aacdecoder_lib.h>
21
22 #include "libavutil/channel_layout.h"
23 #include "libavutil/common.h"
24 #include "libavutil/opt.h"
25 #include "avcodec.h"
26 #include "internal.h"
27
28 #ifdef AACDECODER_LIB_VL0
29 #define FDKDEC_VER_AT_LEAST(vl0, vl1) \
30 ((AACDECODER_LIB_VL0 > vl0) || \
31 (AACDECODER_LIB_VL0 == vl0 && AACDECODER_LIB_VL1 >= vl1))
32 #else
33 #define FDKDEC_VER_AT_LEAST(vl0, vl1) 0
34 #endif
35
36 #if !FDKDEC_VER_AT_LEAST(2, 5) // < 2.5.10
37 #define AAC_PCM_MAX_OUTPUT_CHANNELS AAC_PCM_OUTPUT_CHANNELS
38 #endif
39
40 enum ConcealMethod {
41 CONCEAL_METHOD_SPECTRAL_MUTING = 0,
42 CONCEAL_METHOD_NOISE_SUBSTITUTION = 1,
43 CONCEAL_METHOD_ENERGY_INTERPOLATION = 2,
44 CONCEAL_METHOD_NB,
45 };
46
47 typedef struct FDKAACDecContext {
48 const AVClass *class;
49 HANDLE_AACDECODER handle;
50 uint8_t *decoder_buffer;
51 int decoder_buffer_size;
52 uint8_t *anc_buffer;
53 int conceal_method;
54 int drc_level;
55 int drc_boost;
56 int drc_heavy;
57 int drc_effect;
58 int drc_cut;
59 int level_limit;
60 } FDKAACDecContext;
61
62
63 #define DMX_ANC_BUFFSIZE 128
64 #define DECODER_MAX_CHANNELS 8
65 #define DECODER_BUFFSIZE 2048 * sizeof(INT_PCM)
66
67 #define OFFSET(x) offsetof(FDKAACDecContext, x)
68 #define AD AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_DECODING_PARAM
69 static const AVOption fdk_aac_dec_options[] = {
70 { "conceal", "Error concealment method", OFFSET(conceal_method), AV_OPT_TYPE_INT, { .i64 = CONCEAL_METHOD_NOISE_SUBSTITUTION }, CONCEAL_METHOD_SPECTRAL_MUTING, CONCEAL_METHOD_NB - 1, AD, "conceal" },
71 { "spectral", "Spectral muting", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_SPECTRAL_MUTING }, INT_MIN, INT_MAX, AD, "conceal" },
72 { "noise", "Noise Substitution", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_NOISE_SUBSTITUTION }, INT_MIN, INT_MAX, AD, "conceal" },
73 { "energy", "Energy Interpolation", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_ENERGY_INTERPOLATION }, INT_MIN, INT_MAX, AD, "conceal" },
74 { "drc_boost", "Dynamic Range Control: boost, where [0] is none and [127] is max boost",
75 OFFSET(drc_boost), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
76 { "drc_cut", "Dynamic Range Control: attenuation factor, where [0] is none and [127] is max compression",
77 OFFSET(drc_cut), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
78 { "drc_level", "Dynamic Range Control: reference level, quantized to 0.25dB steps where [0] is 0dB and [127] is -31.75dB",
79 OFFSET(drc_level), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 127, AD, NULL },
80 { "drc_heavy", "Dynamic Range Control: heavy compression, where [1] is on (RF mode) and [0] is off",
81 OFFSET(drc_heavy), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 1, AD, NULL },
82 #if FDKDEC_VER_AT_LEAST(2, 5) // 2.5.10
83 { "level_limit", "Signal level limiting", OFFSET(level_limit), AV_OPT_TYPE_INT, { .i64 = 0 }, -1, 1, AD },
84 #endif
85 #if FDKDEC_VER_AT_LEAST(3, 0) // 3.0.0
86 { "drc_effect","Dynamic Range Control: effect type, where e.g. [0] is none and [6] is general",
87 OFFSET(drc_effect), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 8, AD, NULL },
88 #endif
89 { NULL }
90 };
91
92 static const AVClass fdk_aac_dec_class = {
93 .class_name = "libfdk-aac decoder",
94 .item_name = av_default_item_name,
95 .option = fdk_aac_dec_options,
96 .version = LIBAVUTIL_VERSION_INT,
97 };
98
get_stream_info(AVCodecContext * avctx)99 static int get_stream_info(AVCodecContext *avctx)
100 {
101 FDKAACDecContext *s = avctx->priv_data;
102 CStreamInfo *info = aacDecoder_GetStreamInfo(s->handle);
103 int channel_counts[0x24] = { 0 };
104 int i, ch_error = 0;
105 uint64_t ch_layout = 0;
106
107 if (!info) {
108 av_log(avctx, AV_LOG_ERROR, "Unable to get stream info\n");
109 return AVERROR_UNKNOWN;
110 }
111
112 if (info->sampleRate <= 0) {
113 av_log(avctx, AV_LOG_ERROR, "Stream info not initialized\n");
114 return AVERROR_UNKNOWN;
115 }
116 avctx->sample_rate = info->sampleRate;
117 avctx->frame_size = info->frameSize;
118
119 for (i = 0; i < info->numChannels; i++) {
120 AUDIO_CHANNEL_TYPE ctype = info->pChannelType[i];
121 if (ctype <= ACT_NONE || ctype >= FF_ARRAY_ELEMS(channel_counts)) {
122 av_log(avctx, AV_LOG_WARNING, "unknown channel type\n");
123 break;
124 }
125 channel_counts[ctype]++;
126 }
127 av_log(avctx, AV_LOG_DEBUG,
128 "%d channels - front:%d side:%d back:%d lfe:%d top:%d\n",
129 info->numChannels,
130 channel_counts[ACT_FRONT], channel_counts[ACT_SIDE],
131 channel_counts[ACT_BACK], channel_counts[ACT_LFE],
132 channel_counts[ACT_FRONT_TOP] + channel_counts[ACT_SIDE_TOP] +
133 channel_counts[ACT_BACK_TOP] + channel_counts[ACT_TOP]);
134
135 switch (channel_counts[ACT_FRONT]) {
136 case 4:
137 ch_layout |= AV_CH_LAYOUT_STEREO | AV_CH_FRONT_LEFT_OF_CENTER |
138 AV_CH_FRONT_RIGHT_OF_CENTER;
139 break;
140 case 3:
141 ch_layout |= AV_CH_LAYOUT_STEREO | AV_CH_FRONT_CENTER;
142 break;
143 case 2:
144 ch_layout |= AV_CH_LAYOUT_STEREO;
145 break;
146 case 1:
147 ch_layout |= AV_CH_FRONT_CENTER;
148 break;
149 default:
150 av_log(avctx, AV_LOG_WARNING,
151 "unsupported number of front channels: %d\n",
152 channel_counts[ACT_FRONT]);
153 ch_error = 1;
154 break;
155 }
156 if (channel_counts[ACT_SIDE] > 0) {
157 if (channel_counts[ACT_SIDE] == 2) {
158 ch_layout |= AV_CH_SIDE_LEFT | AV_CH_SIDE_RIGHT;
159 } else {
160 av_log(avctx, AV_LOG_WARNING,
161 "unsupported number of side channels: %d\n",
162 channel_counts[ACT_SIDE]);
163 ch_error = 1;
164 }
165 }
166 if (channel_counts[ACT_BACK] > 0) {
167 switch (channel_counts[ACT_BACK]) {
168 case 3:
169 ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT | AV_CH_BACK_CENTER;
170 break;
171 case 2:
172 ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT;
173 break;
174 case 1:
175 ch_layout |= AV_CH_BACK_CENTER;
176 break;
177 default:
178 av_log(avctx, AV_LOG_WARNING,
179 "unsupported number of back channels: %d\n",
180 channel_counts[ACT_BACK]);
181 ch_error = 1;
182 break;
183 }
184 }
185 if (channel_counts[ACT_LFE] > 0) {
186 if (channel_counts[ACT_LFE] == 1) {
187 ch_layout |= AV_CH_LOW_FREQUENCY;
188 } else {
189 av_log(avctx, AV_LOG_WARNING,
190 "unsupported number of LFE channels: %d\n",
191 channel_counts[ACT_LFE]);
192 ch_error = 1;
193 }
194 }
195 if (!ch_error &&
196 av_get_channel_layout_nb_channels(ch_layout) != info->numChannels) {
197 av_log(avctx, AV_LOG_WARNING, "unsupported channel configuration\n");
198 ch_error = 1;
199 }
200 if (ch_error)
201 avctx->channel_layout = 0;
202 else
203 avctx->channel_layout = ch_layout;
204
205 avctx->channels = info->numChannels;
206
207 return 0;
208 }
209
fdk_aac_decode_close(AVCodecContext * avctx)210 static av_cold int fdk_aac_decode_close(AVCodecContext *avctx)
211 {
212 FDKAACDecContext *s = avctx->priv_data;
213
214 if (s->handle)
215 aacDecoder_Close(s->handle);
216 av_freep(&s->decoder_buffer);
217 av_freep(&s->anc_buffer);
218
219 return 0;
220 }
221
fdk_aac_decode_init(AVCodecContext * avctx)222 static av_cold int fdk_aac_decode_init(AVCodecContext *avctx)
223 {
224 FDKAACDecContext *s = avctx->priv_data;
225 AAC_DECODER_ERROR err;
226
227 s->handle = aacDecoder_Open(avctx->extradata_size ? TT_MP4_RAW : TT_MP4_ADTS, 1);
228 if (!s->handle) {
229 av_log(avctx, AV_LOG_ERROR, "Error opening decoder\n");
230 return AVERROR_UNKNOWN;
231 }
232
233 if (avctx->extradata_size) {
234 if ((err = aacDecoder_ConfigRaw(s->handle, &avctx->extradata,
235 &avctx->extradata_size)) != AAC_DEC_OK) {
236 av_log(avctx, AV_LOG_ERROR, "Unable to set extradata\n");
237 return AVERROR_INVALIDDATA;
238 }
239 }
240
241 if ((err = aacDecoder_SetParam(s->handle, AAC_CONCEAL_METHOD,
242 s->conceal_method)) != AAC_DEC_OK) {
243 av_log(avctx, AV_LOG_ERROR, "Unable to set error concealment method\n");
244 return AVERROR_UNKNOWN;
245 }
246
247 if (avctx->request_channel_layout > 0 &&
248 avctx->request_channel_layout != AV_CH_LAYOUT_NATIVE) {
249 int downmix_channels = -1;
250
251 switch (avctx->request_channel_layout) {
252 case AV_CH_LAYOUT_STEREO:
253 case AV_CH_LAYOUT_STEREO_DOWNMIX:
254 downmix_channels = 2;
255 break;
256 case AV_CH_LAYOUT_MONO:
257 downmix_channels = 1;
258 break;
259 default:
260 av_log(avctx, AV_LOG_WARNING, "Invalid request_channel_layout\n");
261 break;
262 }
263
264 if (downmix_channels != -1) {
265 if (aacDecoder_SetParam(s->handle, AAC_PCM_MAX_OUTPUT_CHANNELS,
266 downmix_channels) != AAC_DEC_OK) {
267 av_log(avctx, AV_LOG_WARNING, "Unable to set output channels in the decoder\n");
268 } else {
269 s->anc_buffer = av_malloc(DMX_ANC_BUFFSIZE);
270 if (!s->anc_buffer) {
271 av_log(avctx, AV_LOG_ERROR, "Unable to allocate ancillary buffer for the decoder\n");
272 return AVERROR(ENOMEM);
273 }
274 if (aacDecoder_AncDataInit(s->handle, s->anc_buffer, DMX_ANC_BUFFSIZE)) {
275 av_log(avctx, AV_LOG_ERROR, "Unable to register downmix ancillary buffer in the decoder\n");
276 return AVERROR_UNKNOWN;
277 }
278 }
279 }
280 }
281
282 if (s->drc_boost != -1) {
283 if (aacDecoder_SetParam(s->handle, AAC_DRC_BOOST_FACTOR, s->drc_boost) != AAC_DEC_OK) {
284 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC boost factor in the decoder\n");
285 return AVERROR_UNKNOWN;
286 }
287 }
288
289 if (s->drc_cut != -1) {
290 if (aacDecoder_SetParam(s->handle, AAC_DRC_ATTENUATION_FACTOR, s->drc_cut) != AAC_DEC_OK) {
291 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC attenuation factor in the decoder\n");
292 return AVERROR_UNKNOWN;
293 }
294 }
295
296 if (s->drc_level != -1) {
297 if (aacDecoder_SetParam(s->handle, AAC_DRC_REFERENCE_LEVEL, s->drc_level) != AAC_DEC_OK) {
298 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC reference level in the decoder\n");
299 return AVERROR_UNKNOWN;
300 }
301 }
302
303 if (s->drc_heavy != -1) {
304 if (aacDecoder_SetParam(s->handle, AAC_DRC_HEAVY_COMPRESSION, s->drc_heavy) != AAC_DEC_OK) {
305 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC heavy compression in the decoder\n");
306 return AVERROR_UNKNOWN;
307 }
308 }
309
310 #if FDKDEC_VER_AT_LEAST(2, 5) // 2.5.10
311 if (aacDecoder_SetParam(s->handle, AAC_PCM_LIMITER_ENABLE, s->level_limit) != AAC_DEC_OK) {
312 av_log(avctx, AV_LOG_ERROR, "Unable to set in signal level limiting in the decoder\n");
313 return AVERROR_UNKNOWN;
314 }
315 #endif
316
317 #if FDKDEC_VER_AT_LEAST(3, 0) // 3.0.0
318 if (s->drc_effect != -1) {
319 if (aacDecoder_SetParam(s->handle, AAC_UNIDRC_SET_EFFECT, s->drc_effect) != AAC_DEC_OK) {
320 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC effect type in the decoder\n");
321 return AVERROR_UNKNOWN;
322 }
323 }
324 #endif
325
326 avctx->sample_fmt = AV_SAMPLE_FMT_S16;
327
328 s->decoder_buffer_size = DECODER_BUFFSIZE * DECODER_MAX_CHANNELS;
329 s->decoder_buffer = av_malloc(s->decoder_buffer_size);
330 if (!s->decoder_buffer)
331 return AVERROR(ENOMEM);
332
333 return 0;
334 }
335
fdk_aac_decode_frame(AVCodecContext * avctx,void * data,int * got_frame_ptr,AVPacket * avpkt)336 static int fdk_aac_decode_frame(AVCodecContext *avctx, void *data,
337 int *got_frame_ptr, AVPacket *avpkt)
338 {
339 FDKAACDecContext *s = avctx->priv_data;
340 AVFrame *frame = data;
341 int ret;
342 AAC_DECODER_ERROR err;
343 UINT valid = avpkt->size;
344
345 err = aacDecoder_Fill(s->handle, &avpkt->data, &avpkt->size, &valid);
346 if (err != AAC_DEC_OK) {
347 av_log(avctx, AV_LOG_ERROR, "aacDecoder_Fill() failed: %x\n", err);
348 return AVERROR_INVALIDDATA;
349 }
350
351 err = aacDecoder_DecodeFrame(s->handle, (INT_PCM *) s->decoder_buffer, s->decoder_buffer_size / sizeof(INT_PCM), 0);
352 if (err == AAC_DEC_NOT_ENOUGH_BITS) {
353 ret = avpkt->size - valid;
354 goto end;
355 }
356 if (err != AAC_DEC_OK) {
357 av_log(avctx, AV_LOG_ERROR,
358 "aacDecoder_DecodeFrame() failed: %x\n", err);
359 ret = AVERROR_UNKNOWN;
360 goto end;
361 }
362
363 if ((ret = get_stream_info(avctx)) < 0)
364 goto end;
365 frame->nb_samples = avctx->frame_size;
366
367 if ((ret = ff_get_buffer(avctx, frame, 0)) < 0)
368 goto end;
369
370 memcpy(frame->extended_data[0], s->decoder_buffer,
371 avctx->channels * avctx->frame_size *
372 av_get_bytes_per_sample(avctx->sample_fmt));
373
374 *got_frame_ptr = 1;
375 ret = avpkt->size - valid;
376
377 end:
378 return ret;
379 }
380
fdk_aac_decode_flush(AVCodecContext * avctx)381 static av_cold void fdk_aac_decode_flush(AVCodecContext *avctx)
382 {
383 FDKAACDecContext *s = avctx->priv_data;
384 AAC_DECODER_ERROR err;
385
386 if (!s->handle)
387 return;
388
389 if ((err = aacDecoder_SetParam(s->handle,
390 AAC_TPDEC_CLEAR_BUFFER, 1)) != AAC_DEC_OK)
391 av_log(avctx, AV_LOG_WARNING, "failed to clear buffer when flushing\n");
392 }
393
394 AVCodec ff_libfdk_aac_decoder = {
395 .name = "libfdk_aac",
396 .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
397 .type = AVMEDIA_TYPE_AUDIO,
398 .id = AV_CODEC_ID_AAC,
399 .priv_data_size = sizeof(FDKAACDecContext),
400 .init = fdk_aac_decode_init,
401 .decode = fdk_aac_decode_frame,
402 .close = fdk_aac_decode_close,
403 .flush = fdk_aac_decode_flush,
404 .capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_CHANNEL_CONF,
405 .priv_class = &fdk_aac_dec_class,
406 .caps_internal = FF_CODEC_CAP_INIT_THREADSAFE |
407 FF_CODEC_CAP_INIT_CLEANUP,
408 .wrapper_name = "libfdk",
409 };
410