c3d3b70fc939d100bcd6c033d48bb78bbb7e731f
[libav.git] / libavcodec / libfdk-aacdec.c
1 /*
2 * AAC decoder wrapper
3 * Copyright (c) 2012 Martin Storsjo
4 *
5 * This file is part of Libav.
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 #define FDKDEC_VER_AT_LEAST(vl0, vl1) \
29 (defined(AACDECODER_LIB_VL0) && \
30 ((AACDECODER_LIB_VL0 > vl0) || \
31 (AACDECODER_LIB_VL0 == vl0 && AACDECODER_LIB_VL1 >= vl1)))
32
33 #if !FDKDEC_VER_AT_LEAST(2, 5) // < 2.5.10
34 #define AAC_PCM_MAX_OUTPUT_CHANNELS AAC_PCM_OUTPUT_CHANNELS
35 #endif
36
37 enum ConcealMethod {
38 CONCEAL_METHOD_SPECTRAL_MUTING = 0,
39 CONCEAL_METHOD_NOISE_SUBSTITUTION = 1,
40 CONCEAL_METHOD_ENERGY_INTERPOLATION = 2,
41 CONCEAL_METHOD_NB,
42 };
43
44 typedef struct FDKAACDecContext {
45 const AVClass *class;
46 HANDLE_AACDECODER handle;
47 uint8_t *decoder_buffer;
48 int decoder_buffer_size;
49 uint8_t *anc_buffer;
50 enum ConcealMethod conceal_method;
51 int drc_level;
52 int drc_boost;
53 int drc_heavy;
54 int drc_cut;
55 int level_limit;
56 } FDKAACDecContext;
57
58
59 #define DMX_ANC_BUFFSIZE 128
60 #define DECODER_MAX_CHANNELS 8
61 #define DECODER_BUFFSIZE 2048 * sizeof(INT_PCM)
62
63 #define OFFSET(x) offsetof(FDKAACDecContext, x)
64 #define AD AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_DECODING_PARAM
65 static const AVOption fdk_aac_dec_options[] = {
66 { "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" },
67 { "spectral", "Spectral muting", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_SPECTRAL_MUTING }, INT_MIN, INT_MAX, AD, "conceal" },
68 { "noise", "Noise Substitution", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_NOISE_SUBSTITUTION }, INT_MIN, INT_MAX, AD, "conceal" },
69 { "energy", "Energy Interpolation", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_ENERGY_INTERPOLATION }, INT_MIN, INT_MAX, AD, "conceal" },
70 { "drc_boost", "Dynamic Range Control: boost, where [0] is none and [127] is max boost",
71 OFFSET(drc_boost), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
72 { "drc_cut", "Dynamic Range Control: attenuation factor, where [0] is none and [127] is max compression",
73 OFFSET(drc_cut), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
74 { "drc_level", "Dynamic Range Control: reference level, quantized to 0.25dB steps where [0] is 0dB and [127] is -31.75dB",
75 OFFSET(drc_level), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 127, AD, NULL },
76 { "drc_heavy", "Dynamic Range Control: heavy compression, where [1] is on (RF mode) and [0] is off",
77 OFFSET(drc_heavy), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 1, AD, NULL },
78 #if FDKDEC_VER_AT_LEAST(2, 5) // 2.5.10
79 { "level_limit", "Signal level limiting", OFFSET(level_limit), AV_OPT_TYPE_INT, { .i64 = 0 }, -1, 1, AD },
80 #endif
81 { NULL }
82 };
83
84 static const AVClass fdk_aac_dec_class = {
85 .class_name = "libfdk-aac decoder",
86 .item_name = av_default_item_name,
87 .option = fdk_aac_dec_options,
88 .version = LIBAVUTIL_VERSION_INT
89 };
90
91 static int get_stream_info(AVCodecContext *avctx)
92 {
93 FDKAACDecContext *s = avctx->priv_data;
94 CStreamInfo *info = aacDecoder_GetStreamInfo(s->handle);
95 int channel_counts[0x24] = { 0 };
96 int i, ch_error = 0;
97 uint64_t ch_layout = 0;
98
99 if (!info) {
100 av_log(avctx, AV_LOG_ERROR, "Unable to get stream info\n");
101 return AVERROR_UNKNOWN;
102 }
103
104 if (info->sampleRate <= 0) {
105 av_log(avctx, AV_LOG_ERROR, "Stream info not initialized\n");
106 return AVERROR_UNKNOWN;
107 }
108 avctx->sample_rate = info->sampleRate;
109 avctx->frame_size = info->frameSize;
110
111 for (i = 0; i < info->numChannels; i++) {
112 AUDIO_CHANNEL_TYPE ctype = info->pChannelType[i];
113 if (ctype <= ACT_NONE || ctype >= FF_ARRAY_ELEMS(channel_counts)) {
114 av_log(avctx, AV_LOG_WARNING, "unknown channel type\n");
115 break;
116 }
117 channel_counts[ctype]++;
118 }
119 av_log(avctx, AV_LOG_DEBUG,
120 "%d channels - front:%d side:%d back:%d lfe:%d top:%d\n",
121 info->numChannels,
122 channel_counts[ACT_FRONT], channel_counts[ACT_SIDE],
123 channel_counts[ACT_BACK], channel_counts[ACT_LFE],
124 channel_counts[ACT_FRONT_TOP] + channel_counts[ACT_SIDE_TOP] +
125 channel_counts[ACT_BACK_TOP] + channel_counts[ACT_TOP]);
126
127 switch (channel_counts[ACT_FRONT]) {
128 case 4:
129 ch_layout |= AV_CH_LAYOUT_STEREO | AV_CH_FRONT_LEFT_OF_CENTER |
130 AV_CH_FRONT_RIGHT_OF_CENTER;
131 break;
132 case 3:
133 ch_layout |= AV_CH_LAYOUT_STEREO | AV_CH_FRONT_CENTER;
134 break;
135 case 2:
136 ch_layout |= AV_CH_LAYOUT_STEREO;
137 break;
138 case 1:
139 ch_layout |= AV_CH_FRONT_CENTER;
140 break;
141 default:
142 av_log(avctx, AV_LOG_WARNING,
143 "unsupported number of front channels: %d\n",
144 channel_counts[ACT_FRONT]);
145 ch_error = 1;
146 break;
147 }
148 if (channel_counts[ACT_SIDE] > 0) {
149 if (channel_counts[ACT_SIDE] == 2) {
150 ch_layout |= AV_CH_SIDE_LEFT | AV_CH_SIDE_RIGHT;
151 } else {
152 av_log(avctx, AV_LOG_WARNING,
153 "unsupported number of side channels: %d\n",
154 channel_counts[ACT_SIDE]);
155 ch_error = 1;
156 }
157 }
158 if (channel_counts[ACT_BACK] > 0) {
159 switch (channel_counts[ACT_BACK]) {
160 case 3:
161 ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT | AV_CH_BACK_CENTER;
162 break;
163 case 2:
164 ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT;
165 break;
166 case 1:
167 ch_layout |= AV_CH_BACK_CENTER;
168 break;
169 default:
170 av_log(avctx, AV_LOG_WARNING,
171 "unsupported number of back channels: %d\n",
172 channel_counts[ACT_BACK]);
173 ch_error = 1;
174 break;
175 }
176 }
177 if (channel_counts[ACT_LFE] > 0) {
178 if (channel_counts[ACT_LFE] == 1) {
179 ch_layout |= AV_CH_LOW_FREQUENCY;
180 } else {
181 av_log(avctx, AV_LOG_WARNING,
182 "unsupported number of LFE channels: %d\n",
183 channel_counts[ACT_LFE]);
184 ch_error = 1;
185 }
186 }
187 if (!ch_error &&
188 av_get_channel_layout_nb_channels(ch_layout) != info->numChannels) {
189 av_log(avctx, AV_LOG_WARNING, "unsupported channel configuration\n");
190 ch_error = 1;
191 }
192 if (ch_error)
193 avctx->channel_layout = 0;
194 else
195 avctx->channel_layout = ch_layout;
196
197 avctx->channels = info->numChannels;
198
199 return 0;
200 }
201
202 static av_cold int fdk_aac_decode_close(AVCodecContext *avctx)
203 {
204 FDKAACDecContext *s = avctx->priv_data;
205
206 if (s->handle)
207 aacDecoder_Close(s->handle);
208 av_free(s->decoder_buffer);
209 av_free(s->anc_buffer);
210
211 return 0;
212 }
213
214 static av_cold int fdk_aac_decode_init(AVCodecContext *avctx)
215 {
216 FDKAACDecContext *s = avctx->priv_data;
217 AAC_DECODER_ERROR err;
218
219 s->handle = aacDecoder_Open(avctx->extradata_size ? TT_MP4_RAW : TT_MP4_ADTS, 1);
220 if (!s->handle) {
221 av_log(avctx, AV_LOG_ERROR, "Error opening decoder\n");
222 return AVERROR_UNKNOWN;
223 }
224
225 if (avctx->extradata_size) {
226 if ((err = aacDecoder_ConfigRaw(s->handle, &avctx->extradata,
227 &avctx->extradata_size)) != AAC_DEC_OK) {
228 av_log(avctx, AV_LOG_ERROR, "Unable to set extradata\n");
229 return AVERROR_INVALIDDATA;
230 }
231 }
232
233 if ((err = aacDecoder_SetParam(s->handle, AAC_CONCEAL_METHOD,
234 s->conceal_method)) != AAC_DEC_OK) {
235 av_log(avctx, AV_LOG_ERROR, "Unable to set error concealment method\n");
236 return AVERROR_UNKNOWN;
237 }
238
239 if (avctx->request_channel_layout > 0 &&
240 avctx->request_channel_layout != AV_CH_LAYOUT_NATIVE) {
241 int downmix_channels = -1;
242
243 switch (avctx->request_channel_layout) {
244 case AV_CH_LAYOUT_STEREO:
245 case AV_CH_LAYOUT_STEREO_DOWNMIX:
246 downmix_channels = 2;
247 break;
248 case AV_CH_LAYOUT_MONO:
249 downmix_channels = 1;
250 break;
251 default:
252 av_log(avctx, AV_LOG_WARNING, "Invalid request_channel_layout\n");
253 break;
254 }
255
256 if (downmix_channels != -1) {
257 if (aacDecoder_SetParam(s->handle, AAC_PCM_MAX_OUTPUT_CHANNELS,
258 downmix_channels) != AAC_DEC_OK) {
259 av_log(avctx, AV_LOG_WARNING, "Unable to set output channels in the decoder\n");
260 } else {
261 s->anc_buffer = av_malloc(DMX_ANC_BUFFSIZE);
262 if (!s->anc_buffer) {
263 av_log(avctx, AV_LOG_ERROR, "Unable to allocate ancillary buffer for the decoder\n");
264 return AVERROR(ENOMEM);
265 }
266 if (aacDecoder_AncDataInit(s->handle, s->anc_buffer, DMX_ANC_BUFFSIZE)) {
267 av_log(avctx, AV_LOG_ERROR, "Unable to register downmix ancillary buffer in the decoder\n");
268 return AVERROR_UNKNOWN;
269 }
270 }
271 }
272 }
273
274 if (s->drc_boost != -1) {
275 if (aacDecoder_SetParam(s->handle, AAC_DRC_BOOST_FACTOR, s->drc_boost) != AAC_DEC_OK) {
276 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC boost factor in the decoder\n");
277 return AVERROR_UNKNOWN;
278 }
279 }
280
281 if (s->drc_cut != -1) {
282 if (aacDecoder_SetParam(s->handle, AAC_DRC_ATTENUATION_FACTOR, s->drc_cut) != AAC_DEC_OK) {
283 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC attenuation factor in the decoder\n");
284 return AVERROR_UNKNOWN;
285 }
286 }
287
288 if (s->drc_level != -1) {
289 if (aacDecoder_SetParam(s->handle, AAC_DRC_REFERENCE_LEVEL, s->drc_level) != AAC_DEC_OK) {
290 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC reference level in the decoder\n");
291 return AVERROR_UNKNOWN;
292 }
293 }
294
295 if (s->drc_heavy != -1) {
296 if (aacDecoder_SetParam(s->handle, AAC_DRC_HEAVY_COMPRESSION, s->drc_heavy) != AAC_DEC_OK) {
297 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC heavy compression in the decoder\n");
298 return AVERROR_UNKNOWN;
299 }
300 }
301
302 #if FDKDEC_VER_AT_LEAST(2, 5) // 2.5.10
303 if (aacDecoder_SetParam(s->handle, AAC_PCM_LIMITER_ENABLE, s->level_limit) != AAC_DEC_OK) {
304 av_log(avctx, AV_LOG_ERROR, "Unable to set in signal level limiting in the decoder\n");
305 return AVERROR_UNKNOWN;
306 }
307 #endif
308
309 avctx->sample_fmt = AV_SAMPLE_FMT_S16;
310
311 s->decoder_buffer_size = DECODER_BUFFSIZE * DECODER_MAX_CHANNELS;
312 s->decoder_buffer = av_malloc(s->decoder_buffer_size);
313 if (!s->decoder_buffer)
314 return AVERROR(ENOMEM);
315
316 return 0;
317 }
318
319 static int fdk_aac_decode_frame(AVCodecContext *avctx, void *data,
320 int *got_frame_ptr, AVPacket *avpkt)
321 {
322 FDKAACDecContext *s = avctx->priv_data;
323 AVFrame *frame = data;
324 int ret;
325 AAC_DECODER_ERROR err;
326 UINT valid = avpkt->size;
327
328 err = aacDecoder_Fill(s->handle, &avpkt->data, &avpkt->size, &valid);
329 if (err != AAC_DEC_OK) {
330 av_log(avctx, AV_LOG_ERROR, "aacDecoder_Fill() failed: %x\n", err);
331 return AVERROR_INVALIDDATA;
332 }
333
334 err = aacDecoder_DecodeFrame(s->handle, (INT_PCM *) s->decoder_buffer, s->decoder_buffer_size / sizeof(INT_PCM), 0);
335 if (err == AAC_DEC_NOT_ENOUGH_BITS) {
336 ret = avpkt->size - valid;
337 goto end;
338 }
339 if (err != AAC_DEC_OK) {
340 av_log(avctx, AV_LOG_ERROR,
341 "aacDecoder_DecodeFrame() failed: %x\n", err);
342 ret = AVERROR_UNKNOWN;
343 goto end;
344 }
345
346 if ((ret = get_stream_info(avctx)) < 0)
347 goto end;
348 frame->nb_samples = avctx->frame_size;
349
350 if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
351 av_log(avctx, AV_LOG_ERROR, "ff_get_buffer() failed\n");
352 goto end;
353 }
354 memcpy(frame->extended_data[0], s->decoder_buffer,
355 avctx->channels * avctx->frame_size *
356 av_get_bytes_per_sample(avctx->sample_fmt));
357
358 *got_frame_ptr = 1;
359 ret = avpkt->size - valid;
360
361 end:
362 return ret;
363 }
364
365 static av_cold void fdk_aac_decode_flush(AVCodecContext *avctx)
366 {
367 FDKAACDecContext *s = avctx->priv_data;
368 AAC_DECODER_ERROR err;
369
370 if (!s->handle)
371 return;
372
373 if ((err = aacDecoder_SetParam(s->handle,
374 AAC_TPDEC_CLEAR_BUFFER, 1)) != AAC_DEC_OK)
375 av_log(avctx, AV_LOG_WARNING, "failed to clear buffer when flushing\n");
376 }
377
378 AVCodec ff_libfdk_aac_decoder = {
379 .name = "libfdk_aac",
380 .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
381 .type = AVMEDIA_TYPE_AUDIO,
382 .id = AV_CODEC_ID_AAC,
383 .priv_data_size = sizeof(FDKAACDecContext),
384 .init = fdk_aac_decode_init,
385 .decode = fdk_aac_decode_frame,
386 .close = fdk_aac_decode_close,
387 .flush = fdk_aac_decode_flush,
388 .capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_CHANNEL_CONF,
389 .priv_class = &fdk_aac_dec_class,
390 .caps_internal = FF_CODEC_CAP_INIT_THREADSAFE |
391 FF_CODEC_CAP_INIT_CLEANUP,
392 .wrapper_name = "libfdk",
393 };