1 /*
2  * AIFF/AIFF-C demuxer
3  * Copyright (c) 2006  Patrick Guimond
4  *
5  * This file is part of FFmpeg.
6  *
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.
11  *
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.
16  *
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
20  */
21 
22 #include "libavutil/intreadwrite.h"
23 #include "libavutil/dict.h"
24 #include "avformat.h"
25 #include "demux.h"
26 #include "internal.h"
27 #include "pcm.h"
28 #include "aiff.h"
29 #include "id3v2.h"
30 #include "mov_chan.h"
31 #include "replaygain.h"
32 
33 #define AIFF                    0
34 #define AIFF_C_VERSION1         0xA2805140
35 
36 typedef struct AIFFInputContext {
37     int64_t data_end;
38     int block_duration;
39 } AIFFInputContext;
40 
aiff_codec_get_id(int bps)41 static enum AVCodecID aiff_codec_get_id(int bps)
42 {
43     if (bps <= 8)
44         return AV_CODEC_ID_PCM_S8;
45     if (bps <= 16)
46         return AV_CODEC_ID_PCM_S16BE;
47     if (bps <= 24)
48         return AV_CODEC_ID_PCM_S24BE;
49     if (bps <= 32)
50         return AV_CODEC_ID_PCM_S32BE;
51 
52     /* bigger than 32 isn't allowed  */
53     return AV_CODEC_ID_NONE;
54 }
55 
56 /* returns the size of the found tag */
get_tag(AVIOContext *pb, uint32_t * tag)57 static int64_t get_tag(AVIOContext *pb, uint32_t * tag)
58 {
59     int64_t size;
60 
61     if (avio_feof(pb))
62         return AVERROR(EIO);
63 
64     *tag = avio_rl32(pb);
65     size = avio_rb32(pb);
66 
67     return size;
68 }
69 
70 /* Metadata string read */
get_meta(AVFormatContext *s, const char *key, int64_t size)71 static void get_meta(AVFormatContext *s, const char *key, int64_t size)
72 {
73     uint8_t *str = NULL;
74 
75     if (size < SIZE_MAX)
76         str = av_malloc(size+1);
77 
78     if (str) {
79         int res = avio_read(s->pb, str, size);
80         if (res < 0){
81             av_free(str);
82             return;
83         }
84         size -= res;
85         str[res] = 0;
86         av_dict_set(&s->metadata, key, str, AV_DICT_DONT_STRDUP_VAL);
87     }
88 
89     avio_skip(s->pb, size);
90 }
91 
92 /* Returns the number of sound data frames or negative on error */
get_aiff_header(AVFormatContext *s, int64_t size, unsigned version)93 static int get_aiff_header(AVFormatContext *s, int64_t size,
94                                     unsigned version)
95 {
96     AVIOContext *pb        = s->pb;
97     AVCodecParameters *par = s->streams[0]->codecpar;
98     AIFFInputContext *aiff = s->priv_data;
99     int exp;
100     uint64_t val;
101     int sample_rate;
102     unsigned int num_frames;
103     int channels;
104 
105     if (size & 1)
106         size++;
107     par->codec_type = AVMEDIA_TYPE_AUDIO;
108     channels = avio_rb16(pb);
109     par->ch_layout.nb_channels = channels;
110     num_frames = avio_rb32(pb);
111     par->bits_per_coded_sample = avio_rb16(pb);
112 
113     exp = avio_rb16(pb) - 16383 - 63;
114     val = avio_rb64(pb);
115     if (exp <-63 || exp >63) {
116         av_log(s, AV_LOG_ERROR, "exp %d is out of range\n", exp);
117         return AVERROR_INVALIDDATA;
118     }
119     if (exp >= 0)
120         sample_rate = val << exp;
121     else
122         sample_rate = (val + (1ULL<<(-exp-1))) >> -exp;
123     if (sample_rate <= 0)
124         return AVERROR_INVALIDDATA;
125 
126     par->sample_rate = sample_rate;
127     if (size < 18)
128         return AVERROR_INVALIDDATA;
129     size -= 18;
130 
131     /* get codec id for AIFF-C */
132     if (size < 4) {
133         version = AIFF;
134     } else if (version == AIFF_C_VERSION1) {
135         par->codec_tag = avio_rl32(pb);
136         par->codec_id  = ff_codec_get_id(ff_codec_aiff_tags, par->codec_tag);
137         if (par->codec_id == AV_CODEC_ID_NONE)
138             avpriv_request_sample(s, "unknown or unsupported codec tag: %s",
139                                   av_fourcc2str(par->codec_tag));
140         size -= 4;
141     }
142 
143     if (version != AIFF_C_VERSION1 || par->codec_id == AV_CODEC_ID_PCM_S16BE) {
144         par->codec_id = aiff_codec_get_id(par->bits_per_coded_sample);
145         par->bits_per_coded_sample = av_get_bits_per_sample(par->codec_id);
146         aiff->block_duration = 1;
147     } else {
148         switch (par->codec_id) {
149         case AV_CODEC_ID_PCM_F32BE:
150         case AV_CODEC_ID_PCM_F64BE:
151         case AV_CODEC_ID_PCM_S16LE:
152         case AV_CODEC_ID_PCM_ALAW:
153         case AV_CODEC_ID_PCM_MULAW:
154             aiff->block_duration = 1;
155             break;
156         case AV_CODEC_ID_ADPCM_IMA_QT:
157             par->block_align = 34 * channels;
158             break;
159         case AV_CODEC_ID_MACE3:
160             par->block_align = 2 * channels;
161             break;
162         case AV_CODEC_ID_ADPCM_G726LE:
163             par->bits_per_coded_sample = 5;
164         case AV_CODEC_ID_ADPCM_IMA_WS:
165         case AV_CODEC_ID_ADPCM_G722:
166         case AV_CODEC_ID_MACE6:
167         case AV_CODEC_ID_SDX2_DPCM:
168             par->block_align = 1 * channels;
169             break;
170         case AV_CODEC_ID_GSM:
171             par->block_align = 33;
172             break;
173         default:
174             aiff->block_duration = 1;
175             break;
176         }
177         if (par->block_align > 0)
178             aiff->block_duration = av_get_audio_frame_duration2(par,
179                                                                 par->block_align);
180     }
181 
182     /* Block align needs to be computed in all cases, as the definition
183      * is specific to applications -> here we use the WAVE format definition */
184     if (!par->block_align)
185         par->block_align = (av_get_bits_per_sample(par->codec_id) * channels) >> 3;
186 
187     if (aiff->block_duration) {
188         par->bit_rate = av_rescale(par->sample_rate, par->block_align * 8LL,
189                                    aiff->block_duration);
190         if (par->bit_rate < 0)
191             par->bit_rate = 0;
192     }
193 
194     /* Chunk is over */
195     if (size)
196         avio_skip(pb, size);
197 
198     return num_frames;
199 }
200 
aiff_probe(const AVProbeData *p)201 static int aiff_probe(const AVProbeData *p)
202 {
203     /* check file header */
204     if (p->buf[0] == 'F' && p->buf[1] == 'O' &&
205         p->buf[2] == 'R' && p->buf[3] == 'M' &&
206         p->buf[8] == 'A' && p->buf[9] == 'I' &&
207         p->buf[10] == 'F' && (p->buf[11] == 'F' || p->buf[11] == 'C'))
208         return AVPROBE_SCORE_MAX;
209     else
210         return 0;
211 }
212 
213 /* aiff input */
aiff_read_header(AVFormatContext *s)214 static int aiff_read_header(AVFormatContext *s)
215 {
216     int ret;
217     int64_t filesize, size;
218     int64_t offset = 0, position;
219     uint32_t tag;
220     unsigned version = AIFF_C_VERSION1;
221     AVIOContext *pb = s->pb;
222     AVStream * st;
223     AIFFInputContext *aiff = s->priv_data;
224     ID3v2ExtraMeta *id3v2_extra_meta;
225 
226     /* check FORM header */
227     filesize = get_tag(pb, &tag);
228     if (filesize < 4 || tag != MKTAG('F', 'O', 'R', 'M'))
229         return AVERROR_INVALIDDATA;
230 
231     /* AIFF data type */
232     tag = avio_rl32(pb);
233     if (tag == MKTAG('A', 'I', 'F', 'F'))       /* Got an AIFF file */
234         version = AIFF;
235     else if (tag != MKTAG('A', 'I', 'F', 'C'))  /* An AIFF-C file then */
236         return AVERROR_INVALIDDATA;
237 
238     filesize -= 4;
239 
240     st = avformat_new_stream(s, NULL);
241     if (!st)
242         return AVERROR(ENOMEM);
243 
244     while (filesize > 0) {
245         /* parse different chunks */
246         size = get_tag(pb, &tag);
247 
248         if (size == AVERROR_EOF && offset > 0 && st->codecpar->block_align) {
249             av_log(s, AV_LOG_WARNING, "header parser hit EOF\n");
250             goto got_sound;
251         }
252         if (size < 0)
253             return size;
254 
255         filesize -= size + 8;
256 
257         switch (tag) {
258         case MKTAG('C', 'O', 'M', 'M'):     /* Common chunk */
259             /* Then for the complete header info */
260             st->nb_frames = get_aiff_header(s, size, version);
261             if (st->nb_frames < 0)
262                 return st->nb_frames;
263             if (offset > 0) // COMM is after SSND
264                 goto got_sound;
265             break;
266         case MKTAG('I', 'D', '3', ' '):
267             position = avio_tell(pb);
268             ff_id3v2_read(s, ID3v2_DEFAULT_MAGIC, &id3v2_extra_meta, size);
269             if (id3v2_extra_meta)
270                 if ((ret = ff_id3v2_parse_apic(s, id3v2_extra_meta)) < 0 ||
271                     (ret = ff_id3v2_parse_chapters(s, id3v2_extra_meta)) < 0) {
272                     ff_id3v2_free_extra_meta(&id3v2_extra_meta);
273                     return ret;
274                 }
275             ff_id3v2_free_extra_meta(&id3v2_extra_meta);
276             if (position + size > avio_tell(pb))
277                 avio_skip(pb, position + size - avio_tell(pb));
278             break;
279         case MKTAG('F', 'V', 'E', 'R'):     /* Version chunk */
280             version = avio_rb32(pb);
281             break;
282         case MKTAG('N', 'A', 'M', 'E'):     /* Sample name chunk */
283             get_meta(s, "title"    , size);
284             break;
285         case MKTAG('A', 'U', 'T', 'H'):     /* Author chunk */
286             get_meta(s, "author"   , size);
287             break;
288         case MKTAG('(', 'c', ')', ' '):     /* Copyright chunk */
289             get_meta(s, "copyright", size);
290             break;
291         case MKTAG('A', 'N', 'N', 'O'):     /* Annotation chunk */
292             get_meta(s, "comment"  , size);
293             break;
294         case MKTAG('S', 'S', 'N', 'D'):     /* Sampled sound chunk */
295             if (size < 8)
296                 return AVERROR_INVALIDDATA;
297             aiff->data_end = avio_tell(pb) + size;
298             offset = avio_rb32(pb);      /* Offset of sound data */
299             avio_rb32(pb);               /* BlockSize... don't care */
300             offset += avio_tell(pb);    /* Compute absolute data offset */
301             if (st->codecpar->block_align && !(pb->seekable & AVIO_SEEKABLE_NORMAL))    /* Assume COMM already parsed */
302                 goto got_sound;
303             if (!(pb->seekable & AVIO_SEEKABLE_NORMAL)) {
304                 av_log(s, AV_LOG_ERROR, "file is not seekable\n");
305                 return -1;
306             }
307             avio_skip(pb, size - 8);
308             break;
309         case MKTAG('w', 'a', 'v', 'e'):
310             if ((uint64_t)size > (1<<30))
311                 return AVERROR_INVALIDDATA;
312             if ((ret = ff_get_extradata(s, st->codecpar, pb, size)) < 0)
313                 return ret;
314             if (   (st->codecpar->codec_id == AV_CODEC_ID_QDMC || st->codecpar->codec_id == AV_CODEC_ID_QDM2)
315                 && size>=12*4 && !st->codecpar->block_align) {
316                 st->codecpar->block_align = AV_RB32(st->codecpar->extradata+11*4);
317                 aiff->block_duration = AV_RB32(st->codecpar->extradata+9*4);
318             } else if (st->codecpar->codec_id == AV_CODEC_ID_QCELP) {
319                 char rate = 0;
320                 if (size >= 25)
321                     rate = st->codecpar->extradata[24];
322                 switch (rate) {
323                 case 'H': // RATE_HALF
324                     st->codecpar->block_align = 17;
325                     break;
326                 case 'F': // RATE_FULL
327                 default:
328                     st->codecpar->block_align = 35;
329                 }
330                 aiff->block_duration = 160;
331                 st->codecpar->bit_rate = (int64_t)st->codecpar->sample_rate * (st->codecpar->block_align << 3) /
332                                          aiff->block_duration;
333             }
334             break;
335         case MKTAG('C','H','A','N'):
336             if ((ret = ff_mov_read_chan(s, pb, st, size)) < 0)
337                 return ret;
338             break;
339         case MKTAG('A','P','C','M'): /* XA ADPCM compressed sound chunk */
340             st->codecpar->codec_id = AV_CODEC_ID_ADPCM_XA;
341             aiff->data_end = avio_tell(pb) + size;
342             offset = avio_tell(pb) + 8;
343             /* This field is unknown and its data seems to be irrelevant */
344             avio_rb32(pb);
345             st->codecpar->block_align = avio_rb32(pb);
346 
347             goto got_sound;
348             break;
349         case 0:
350             if (offset > 0 && st->codecpar->block_align) // COMM && SSND
351                 goto got_sound;
352         default: /* Jump */
353             avio_skip(pb, size);
354         }
355 
356         /* Skip required padding byte for odd-sized chunks. */
357         if (size & 1) {
358             filesize--;
359             avio_skip(pb, 1);
360         }
361     }
362 
363     ret = ff_replaygain_export(st, s->metadata);
364     if (ret < 0)
365         return ret;
366 
367 got_sound:
368     if (!st->codecpar->block_align && st->codecpar->codec_id == AV_CODEC_ID_QCELP) {
369         av_log(s, AV_LOG_WARNING, "qcelp without wave chunk, assuming full rate\n");
370         st->codecpar->block_align = 35;
371     } else if (st->codecpar->block_align <= 0) {
372         av_log(s, AV_LOG_ERROR, "could not find COMM tag or invalid block_align value\n");
373         return AVERROR_INVALIDDATA;
374     }
375     if (aiff->block_duration < 0)
376         return AVERROR_INVALIDDATA;
377 
378     /* Now positioned, get the sound data start and end */
379     avpriv_set_pts_info(st, 64, 1, st->codecpar->sample_rate);
380     st->start_time = 0;
381     st->duration = st->nb_frames * aiff->block_duration;
382 
383     /* Position the stream at the first block */
384     avio_seek(pb, offset, SEEK_SET);
385 
386     return 0;
387 }
388 
389 #define MAX_SIZE 4096
390 
aiff_read_packet(AVFormatContext *s, AVPacket *pkt)391 static int aiff_read_packet(AVFormatContext *s,
392                             AVPacket *pkt)
393 {
394     AVStream *st = s->streams[0];
395     AIFFInputContext *aiff = s->priv_data;
396     int64_t max_size;
397     int res, size;
398 
399     /* calculate size of remaining data */
400     max_size = aiff->data_end - avio_tell(s->pb);
401     if (max_size <= 0)
402         return AVERROR_EOF;
403 
404     if (!st->codecpar->block_align) {
405         av_log(s, AV_LOG_ERROR, "block_align not set\n");
406         return AVERROR_INVALIDDATA;
407     }
408 
409     /* Now for that packet */
410     switch (st->codecpar->codec_id) {
411     case AV_CODEC_ID_ADPCM_IMA_QT:
412     case AV_CODEC_ID_GSM:
413     case AV_CODEC_ID_QDM2:
414     case AV_CODEC_ID_QCELP:
415         size = st->codecpar->block_align;
416         break;
417     default:
418         size = st->codecpar->block_align ? (MAX_SIZE / st->codecpar->block_align) * st->codecpar->block_align : MAX_SIZE;
419         if (!size)
420             return AVERROR_INVALIDDATA;
421     }
422     size = FFMIN(max_size, size);
423     res = av_get_packet(s->pb, pkt, size);
424     if (res < 0)
425         return res;
426 
427     if (size >= st->codecpar->block_align)
428         pkt->flags &= ~AV_PKT_FLAG_CORRUPT;
429     /* Only one stream in an AIFF file */
430     pkt->stream_index = 0;
431     pkt->duration     = (res / st->codecpar->block_align) * (int64_t) aiff->block_duration;
432     return 0;
433 }
434 
435 const AVInputFormat ff_aiff_demuxer = {
436     .name           = "aiff",
437     .long_name      = NULL_IF_CONFIG_SMALL("Audio IFF"),
438     .priv_data_size = sizeof(AIFFInputContext),
439     .read_probe     = aiff_probe,
440     .read_header    = aiff_read_header,
441     .read_packet    = aiff_read_packet,
442     .read_seek      = ff_pcm_read_seek,
443     .codec_tag      = ff_aiff_codec_tags_list,
444 };
445