libfdk-aacdec: Allow setting the new dynamic range control effect setting
[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_effect;
55 int drc_cut;
56 int level_limit;
57 } FDKAACDecContext;
58
59
60 #define DMX_ANC_BUFFSIZE 128
61 #define DECODER_MAX_CHANNELS 8
62 #define DECODER_BUFFSIZE 2048 * sizeof(INT_PCM)
63
64 #define OFFSET(x) offsetof(FDKAACDecContext, x)
65 #define AD AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_DECODING_PARAM
66 static const AVOption fdk_aac_dec_options[] = {
67 { "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" },
68 { "spectral", "Spectral muting", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_SPECTRAL_MUTING }, INT_MIN, INT_MAX, AD, "conceal" },
69 { "noise", "Noise Substitution", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_NOISE_SUBSTITUTION }, INT_MIN, INT_MAX, AD, "conceal" },
70 { "energy", "Energy Interpolation", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_ENERGY_INTERPOLATION }, INT_MIN, INT_MAX, AD, "conceal" },
71 { "drc_boost", "Dynamic Range Control: boost, where [0] is none and [127] is max boost",
72 OFFSET(drc_boost), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
73 { "drc_cut", "Dynamic Range Control: attenuation factor, where [0] is none and [127] is max compression",
74 OFFSET(drc_cut), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
75 { "drc_level", "Dynamic Range Control: reference level, quantized to 0.25dB steps where [0] is 0dB and [127] is -31.75dB",
76 OFFSET(drc_level), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 127, AD, NULL },
77 { "drc_heavy", "Dynamic Range Control: heavy compression, where [1] is on (RF mode) and [0] is off",
78 OFFSET(drc_heavy), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 1, AD, NULL },
79 #if FDKDEC_VER_AT_LEAST(2, 5) // 2.5.10
80 { "level_limit", "Signal level limiting", OFFSET(level_limit), AV_OPT_TYPE_INT, { .i64 = 0 }, -1, 1, AD },
81 #endif
82 #if FDKDEC_VER_AT_LEAST(3, 0) // 3.0.0
83 { "drc_effect","Dynamic Range Control: effect type, where e.g. [0] is none and [6] is general",
84 OFFSET(drc_effect), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 8, AD, NULL },
85 #endif
86 { NULL }
87 };
88
89 static const AVClass fdk_aac_dec_class = {
90 .class_name = "libfdk-aac decoder",
91 .item_name = av_default_item_name,
92 .option = fdk_aac_dec_options,
93 .version = LIBAVUTIL_VERSION_INT
94 };
95
96 static int get_stream_info(AVCodecContext *avctx)
97 {
98 FDKAACDecContext *s = avctx->priv_data;
99 CStreamInfo *info = aacDecoder_GetStreamInfo(s->handle);
100 int channel_counts[0x24] = { 0 };
101 int i, ch_error = 0;
102 uint64_t ch_layout = 0;
103
104 if (!info) {
105 av_log(avctx, AV_LOG_ERROR, "Unable to get stream info\n");
106 return AVERROR_UNKNOWN;
107 }
108
109 if (info->sampleRate <= 0) {
110 av_log(avctx, AV_LOG_ERROR, "Stream info not initialized\n");
111 return AVERROR_UNKNOWN;
112 }
113 avctx->sample_rate = info->sampleRate;
114 avctx->frame_size = info->frameSize;
115
116 for (i = 0; i < info->numChannels; i++) {
117 AUDIO_CHANNEL_TYPE ctype = info->pChannelType[i];
118 if (ctype <= ACT_NONE || ctype >= FF_ARRAY_ELEMS(channel_counts)) {
119 av_log(avctx, AV_LOG_WARNING, "unknown channel type\n");
120 break;
121 }
122 channel_counts[ctype]++;
123 }
124 av_log(avctx, AV_LOG_DEBUG,
125 "%d channels - front:%d side:%d back:%d lfe:%d top:%d\n",
126 info->numChannels,
127 channel_counts[ACT_FRONT], channel_counts[ACT_SIDE],
128 channel_counts[ACT_BACK], channel_counts[ACT_LFE],
129 channel_counts[ACT_FRONT_TOP] + channel_counts[ACT_SIDE_TOP] +
130 channel_counts[ACT_BACK_TOP] + channel_counts[ACT_TOP]);
131
132 switch (channel_counts[ACT_FRONT]) {
133 case 4:
134 ch_layout |= AV_CH_LAYOUT_STEREO | AV_CH_FRONT_LEFT_OF_CENTER |
135 AV_CH_FRONT_RIGHT_OF_CENTER;
136 break;
137 case 3:
138 ch_layout |= AV_CH_LAYOUT_STEREO | AV_CH_FRONT_CENTER;
139 break;
140 case 2:
141 ch_layout |= AV_CH_LAYOUT_STEREO;
142 break;
143 case 1:
144 ch_layout |= AV_CH_FRONT_CENTER;
145 break;
146 default:
147 av_log(avctx, AV_LOG_WARNING,
148 "unsupported number of front channels: %d\n",
149 channel_counts[ACT_FRONT]);
150 ch_error = 1;
151 break;
152 }
153 if (channel_counts[ACT_SIDE] > 0) {
154 if (channel_counts[ACT_SIDE] == 2) {
155 ch_layout |= AV_CH_SIDE_LEFT | AV_CH_SIDE_RIGHT;
156 } else {
157 av_log(avctx, AV_LOG_WARNING,
158 "unsupported number of side channels: %d\n",
159 channel_counts[ACT_SIDE]);
160 ch_error = 1;
161 }
162 }
163 if (channel_counts[ACT_BACK] > 0) {
164 switch (channel_counts[ACT_BACK]) {
165 case 3:
166 ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT | AV_CH_BACK_CENTER;
167 break;
168 case 2:
169 ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT;
170 break;
171 case 1:
172 ch_layout |= AV_CH_BACK_CENTER;
173 break;
174 default:
175 av_log(avctx, AV_LOG_WARNING,
176 "unsupported number of back channels: %d\n",
177 channel_counts[ACT_BACK]);
178 ch_error = 1;
179 break;
180 }
181 }
182 if (channel_counts[ACT_LFE] > 0) {
183 if (channel_counts[ACT_LFE] == 1) {
184 ch_layout |= AV_CH_LOW_FREQUENCY;
185 } else {
186 av_log(avctx, AV_LOG_WARNING,
187 "unsupported number of LFE channels: %d\n",
188 channel_counts[ACT_LFE]);
189 ch_error = 1;
190 }
191 }
192 if (!ch_error &&
193 av_get_channel_layout_nb_channels(ch_layout) != info->numChannels) {
194 av_log(avctx, AV_LOG_WARNING, "unsupported channel configuration\n");
195 ch_error = 1;
196 }
197 if (ch_error)
198 avctx->channel_layout = 0;
199 else
200 avctx->channel_layout = ch_layout;
201
202 avctx->channels = info->numChannels;
203
204 return 0;
205 }
206
207 static av_cold int fdk_aac_decode_close(AVCodecContext *avctx)
208 {
209 FDKAACDecContext *s = avctx->priv_data;
210
211 if (s->handle)
212 aacDecoder_Close(s->handle);
213 av_free(s->decoder_buffer);
214 av_free(s->anc_buffer);
215
216 return 0;
217 }
218
219 static av_cold int fdk_aac_decode_init(AVCodecContext *avctx)
220 {
221 FDKAACDecContext *s = avctx->priv_data;
222 AAC_DECODER_ERROR err;
223
224 s->handle = aacDecoder_Open(avctx->extradata_size ? TT_MP4_RAW : TT_MP4_ADTS, 1);
225 if (!s->handle) {
226 av_log(avctx, AV_LOG_ERROR, "Error opening decoder\n");
227 return AVERROR_UNKNOWN;
228 }
229
230 if (avctx->extradata_size) {
231 if ((err = aacDecoder_ConfigRaw(s->handle, &avctx->extradata,
232 &avctx->extradata_size)) != AAC_DEC_OK) {
233 av_log(avctx, AV_LOG_ERROR, "Unable to set extradata\n");
234 return AVERROR_INVALIDDATA;
235 }
236 }
237
238 if ((err = aacDecoder_SetParam(s->handle, AAC_CONCEAL_METHOD,
239 s->conceal_method)) != AAC_DEC_OK) {
240 av_log(avctx, AV_LOG_ERROR, "Unable to set error concealment method\n");
241 return AVERROR_UNKNOWN;
242 }
243
244 if (avctx->request_channel_layout > 0 &&
245 avctx->request_channel_layout != AV_CH_LAYOUT_NATIVE) {
246 int downmix_channels = -1;
247
248 switch (avctx->request_channel_layout) {
249 case AV_CH_LAYOUT_STEREO:
250 case AV_CH_LAYOUT_STEREO_DOWNMIX:
251 downmix_channels = 2;
252 break;
253 case AV_CH_LAYOUT_MONO:
254 downmix_channels = 1;
255 break;
256 default:
257 av_log(avctx, AV_LOG_WARNING, "Invalid request_channel_layout\n");
258 break;
259 }
260
261 if (downmix_channels != -1) {
262 if (aacDecoder_SetParam(s->handle, AAC_PCM_MAX_OUTPUT_CHANNELS,
263 downmix_channels) != AAC_DEC_OK) {
264 av_log(avctx, AV_LOG_WARNING, "Unable to set output channels in the decoder\n");
265 } else {
266 s->anc_buffer = av_malloc(DMX_ANC_BUFFSIZE);
267 if (!s->anc_buffer) {
268 av_log(avctx, AV_LOG_ERROR, "Unable to allocate ancillary buffer for the decoder\n");
269 return AVERROR(ENOMEM);
270 }
271 if (aacDecoder_AncDataInit(s->handle, s->anc_buffer, DMX_ANC_BUFFSIZE)) {
272 av_log(avctx, AV_LOG_ERROR, "Unable to register downmix ancillary buffer in the decoder\n");
273 return AVERROR_UNKNOWN;
274 }
275 }
276 }
277 }
278
279 if (s->drc_boost != -1) {
280 if (aacDecoder_SetParam(s->handle, AAC_DRC_BOOST_FACTOR, s->drc_boost) != AAC_DEC_OK) {
281 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC boost factor in the decoder\n");
282 return AVERROR_UNKNOWN;
283 }
284 }
285
286 if (s->drc_cut != -1) {
287 if (aacDecoder_SetParam(s->handle, AAC_DRC_ATTENUATION_FACTOR, s->drc_cut) != AAC_DEC_OK) {
288 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC attenuation factor in the decoder\n");
289 return AVERROR_UNKNOWN;
290 }
291 }
292
293 if (s->drc_level != -1) {
294 if (aacDecoder_SetParam(s->handle, AAC_DRC_REFERENCE_LEVEL, s->drc_level) != AAC_DEC_OK) {
295 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC reference level in the decoder\n");
296 return AVERROR_UNKNOWN;
297 }
298 }
299
300 if (s->drc_heavy != -1) {
301 if (aacDecoder_SetParam(s->handle, AAC_DRC_HEAVY_COMPRESSION, s->drc_heavy) != AAC_DEC_OK) {
302 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC heavy compression in the decoder\n");
303 return AVERROR_UNKNOWN;
304 }
305 }
306
307 #if FDKDEC_VER_AT_LEAST(2, 5) // 2.5.10
308 if (aacDecoder_SetParam(s->handle, AAC_PCM_LIMITER_ENABLE, s->level_limit) != AAC_DEC_OK) {
309 av_log(avctx, AV_LOG_ERROR, "Unable to set in signal level limiting in the decoder\n");
310 return AVERROR_UNKNOWN;
311 }
312 #endif
313
314 #if FDKDEC_VER_AT_LEAST(3, 0) // 3.0.0
315 if (s->drc_effect != -1) {
316 if (aacDecoder_SetParam(s->handle, AAC_UNIDRC_SET_EFFECT, s->drc_effect) != AAC_DEC_OK) {
317 av_log(avctx, AV_LOG_ERROR, "Unable to set DRC effect type in the decoder\n");
318 return AVERROR_UNKNOWN;
319 }
320 }
321 #endif
322
323 avctx->sample_fmt = AV_SAMPLE_FMT_S16;
324
325 s->decoder_buffer_size = DECODER_BUFFSIZE * DECODER_MAX_CHANNELS;
326 s->decoder_buffer = av_malloc(s->decoder_buffer_size);
327 if (!s->decoder_buffer)
328 return AVERROR(ENOMEM);
329
330 return 0;
331 }
332
333 static int fdk_aac_decode_frame(AVCodecContext *avctx, void *data,
334 int *got_frame_ptr, AVPacket *avpkt)
335 {
336 FDKAACDecContext *s = avctx->priv_data;
337 AVFrame *frame = data;
338 int ret;
339 AAC_DECODER_ERROR err;
340 UINT valid = avpkt->size;
341
342 err = aacDecoder_Fill(s->handle, &avpkt->data, &avpkt->size, &valid);
343 if (err != AAC_DEC_OK) {
344 av_log(avctx, AV_LOG_ERROR, "aacDecoder_Fill() failed: %x\n", err);
345 return AVERROR_INVALIDDATA;
346 }
347
348 err = aacDecoder_DecodeFrame(s->handle, (INT_PCM *) s->decoder_buffer, s->decoder_buffer_size / sizeof(INT_PCM), 0);
349 if (err == AAC_DEC_NOT_ENOUGH_BITS) {
350 ret = avpkt->size - valid;
351 goto end;
352 }
353 if (err != AAC_DEC_OK) {
354 av_log(avctx, AV_LOG_ERROR,
355 "aacDecoder_DecodeFrame() failed: %x\n", err);
356 ret = AVERROR_UNKNOWN;
357 goto end;
358 }
359
360 if ((ret = get_stream_info(avctx)) < 0)
361 goto end;
362 frame->nb_samples = avctx->frame_size;
363
364 if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
365 av_log(avctx, AV_LOG_ERROR, "ff_get_buffer() failed\n");
366 goto end;
367 }
368 memcpy(frame->extended_data[0], s->decoder_buffer,
369 avctx->channels * avctx->frame_size *
370 av_get_bytes_per_sample(avctx->sample_fmt));
371
372 *got_frame_ptr = 1;
373 ret = avpkt->size - valid;
374
375 end:
376 return ret;
377 }
378
379 static av_cold void fdk_aac_decode_flush(AVCodecContext *avctx)
380 {
381 FDKAACDecContext *s = avctx->priv_data;
382 AAC_DECODER_ERROR err;
383
384 if (!s->handle)
385 return;
386
387 if ((err = aacDecoder_SetParam(s->handle,
388 AAC_TPDEC_CLEAR_BUFFER, 1)) != AAC_DEC_OK)
389 av_log(avctx, AV_LOG_WARNING, "failed to clear buffer when flushing\n");
390 }
391
392 AVCodec ff_libfdk_aac_decoder = {
393 .name = "libfdk_aac",
394 .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
395 .type = AVMEDIA_TYPE_AUDIO,
396 .id = AV_CODEC_ID_AAC,
397 .priv_data_size = sizeof(FDKAACDecContext),
398 .init = fdk_aac_decode_init,
399 .decode = fdk_aac_decode_frame,
400 .close = fdk_aac_decode_close,
401 .flush = fdk_aac_decode_flush,
402 .capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_CHANNEL_CONF,
403 .priv_class = &fdk_aac_dec_class,
404 .caps_internal = FF_CODEC_CAP_INIT_THREADSAFE |
405 FF_CODEC_CAP_INIT_CLEANUP,
406 .wrapper_name = "libfdk",
407 };