2 * Core Audio Format demuxer
3 * Copyright (c) 2007 Justin Ruggles
4 * Copyright (c) 2009 Peter Ross
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
25 * Core Audio Format demuxer
31 #include "libavutil/intreadwrite.h"
32 #include "libavutil/dict.h"
36 int bytes_per_packet; ///< bytes in a packet, or 0 if variable
37 int frames_per_packet; ///< frames in a packet, or 0 if variable
38 int64_t num_bytes; ///< total number of bytes in stream
40 int64_t packet_cnt; ///< packet counter
41 int64_t frame_cnt; ///< frame counter
43 int64_t data_start; ///< data start position, in bytes
44 int64_t data_size; ///< raw data size, in bytes
47 static int probe(AVProbeData *p)
49 if (AV_RB32(p->buf) == MKBETAG('c','a','f','f') && AV_RB16(&p->buf[4]) == 1)
50 return AVPROBE_SCORE_MAX;
54 /** Read audio description chunk */
55 static int read_desc_chunk(AVFormatContext *s)
57 AVIOContext *pb = s->pb;
58 CaffContext *caf = s->priv_data;
62 /* new audio stream */
63 st = av_new_stream(s, 0);
65 return AVERROR(ENOMEM);
67 /* parse format description */
68 st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
69 st->codec->sample_rate = av_int2dbl(avio_rb64(pb));
70 st->codec->codec_tag = avio_rb32(pb);
71 flags = avio_rb32(pb);
72 caf->bytes_per_packet = avio_rb32(pb);
73 st->codec->block_align = caf->bytes_per_packet;
74 caf->frames_per_packet = avio_rb32(pb);
75 st->codec->channels = avio_rb32(pb);
76 st->codec->bits_per_coded_sample = avio_rb32(pb);
78 /* calculate bit rate for constant size packets */
79 if (caf->frames_per_packet > 0 && caf->bytes_per_packet > 0) {
80 st->codec->bit_rate = (uint64_t)st->codec->sample_rate * (uint64_t)caf->bytes_per_packet * 8
81 / (uint64_t)caf->frames_per_packet;
83 st->codec->bit_rate = 0;
87 if (st->codec->codec_tag == MKBETAG('l','p','c','m'))
88 st->codec->codec_id = ff_mov_get_lpcm_codec_id(st->codec->bits_per_coded_sample, (flags ^ 0x2) | 0x4);
90 st->codec->codec_id = ff_codec_get_id(ff_codec_caf_tags, st->codec->codec_tag);
94 /** Read magic cookie chunk */
95 static int read_kuki_chunk(AVFormatContext *s, int64_t size)
97 AVIOContext *pb = s->pb;
98 AVStream *st = s->streams[0];
100 if (size < 0 || size > INT_MAX - FF_INPUT_BUFFER_PADDING_SIZE)
103 if (st->codec->codec_id == CODEC_ID_AAC) {
104 /* The magic cookie format for AAC is an mp4 esds atom.
105 The lavc AAC decoder requires the data from the codec specific
106 description as extradata input. */
110 strt = avio_tell(pb);
111 ff_mov_read_esds(s, pb, atom);
112 skip = size - (avio_tell(pb) - strt);
113 if (skip < 0 || !st->codec->extradata ||
114 st->codec->codec_id != CODEC_ID_AAC) {
115 av_log(s, AV_LOG_ERROR, "invalid AAC magic cookie\n");
116 return AVERROR_INVALIDDATA;
119 } else if (st->codec->codec_id == CODEC_ID_ALAC) {
120 #define ALAC_PREAMBLE 12
121 #define ALAC_HEADER 36
122 if (size < ALAC_PREAMBLE + ALAC_HEADER) {
123 av_log(s, AV_LOG_ERROR, "invalid ALAC magic cookie\n");
125 return AVERROR_INVALIDDATA;
127 avio_skip(pb, ALAC_PREAMBLE);
128 st->codec->extradata = av_mallocz(ALAC_HEADER + FF_INPUT_BUFFER_PADDING_SIZE);
129 if (!st->codec->extradata)
130 return AVERROR(ENOMEM);
131 avio_read(pb, st->codec->extradata, ALAC_HEADER);
132 st->codec->extradata_size = ALAC_HEADER;
133 avio_skip(pb, size - ALAC_PREAMBLE - ALAC_HEADER);
135 st->codec->extradata = av_mallocz(size + FF_INPUT_BUFFER_PADDING_SIZE);
136 if (!st->codec->extradata)
137 return AVERROR(ENOMEM);
138 avio_read(pb, st->codec->extradata, size);
139 st->codec->extradata_size = size;
145 /** Read packet table chunk */
146 static int read_pakt_chunk(AVFormatContext *s, int64_t size)
148 AVIOContext *pb = s->pb;
149 AVStream *st = s->streams[0];
150 CaffContext *caf = s->priv_data;
151 int64_t pos = 0, ccount;
154 ccount = avio_tell(pb);
156 num_packets = avio_rb64(pb);
157 if (num_packets < 0 || INT32_MAX / sizeof(AVIndexEntry) < num_packets)
158 return AVERROR_INVALIDDATA;
160 st->nb_frames = avio_rb64(pb); /* valid frames */
161 st->nb_frames += avio_rb32(pb); /* priming frames */
162 st->nb_frames += avio_rb32(pb); /* remainder frames */
165 for (i = 0; i < num_packets; i++) {
166 av_add_index_entry(s->streams[0], pos, st->duration, 0, 0, AVINDEX_KEYFRAME);
167 pos += caf->bytes_per_packet ? caf->bytes_per_packet : ff_mp4_read_descr_len(pb);
168 st->duration += caf->frames_per_packet ? caf->frames_per_packet : ff_mp4_read_descr_len(pb);
171 if (avio_tell(pb) - ccount != size) {
172 av_log(s, AV_LOG_ERROR, "error reading packet table\n");
176 caf->num_bytes = pos;
180 /** Read information chunk */
181 static void read_info_chunk(AVFormatContext *s, int64_t size)
183 AVIOContext *pb = s->pb;
185 unsigned int nb_entries = avio_rb32(pb);
186 for (i = 0; i < nb_entries; i++) {
189 get_strz(pb, key, sizeof(key));
190 get_strz(pb, value, sizeof(value));
191 av_dict_set(&s->metadata, key, value, 0);
195 static int read_header(AVFormatContext *s,
196 AVFormatParameters *ap)
198 AVIOContext *pb = s->pb;
199 CaffContext *caf = s->priv_data;
205 avio_skip(pb, 8); /* magic, version, file flags */
207 /* audio description chunk */
208 if (avio_rb32(pb) != MKBETAG('d','e','s','c')) {
209 av_log(s, AV_LOG_ERROR, "desc chunk not present\n");
210 return AVERROR_INVALIDDATA;
212 size = avio_rb64(pb);
214 return AVERROR_INVALIDDATA;
216 ret = read_desc_chunk(s);
221 /* parse each chunk */
223 while (!url_feof(pb)) {
225 /* stop at data chunk if seeking is not supported or
226 data chunk size is unknown */
227 if (found_data && (caf->data_size < 0 || !pb->seekable))
231 size = avio_rb64(pb);
236 case MKBETAG('d','a','t','a'):
237 avio_skip(pb, 4); /* edit count */
238 caf->data_start = avio_tell(pb);
239 caf->data_size = size < 0 ? -1 : size - 4;
240 if (caf->data_size > 0 && pb->seekable)
241 avio_skip(pb, caf->data_size);
245 /* magic cookie chunk */
246 case MKBETAG('k','u','k','i'):
247 if (read_kuki_chunk(s, size))
248 return AVERROR_INVALIDDATA;
251 /* packet table chunk */
252 case MKBETAG('p','a','k','t'):
253 if (read_pakt_chunk(s, size))
254 return AVERROR_INVALIDDATA;
257 case MKBETAG('i','n','f','o'):
258 read_info_chunk(s, size);
261 case MKBETAG('c','h','a','n'):
263 return AVERROR_INVALIDDATA;
264 ff_mov_read_chan(s, size, st->codec);
268 #define _(x) ((x) >= ' ' ? (x) : ' ')
269 av_log(s, AV_LOG_WARNING, "skipping CAF chunk: %08X (%c%c%c%c), size %"PRId64"\n",
270 tag, _(tag>>24), _((tag>>16)&0xFF), _((tag>>8)&0xFF), _(tag&0xFF), size);
272 case MKBETAG('f','r','e','e'):
274 return AVERROR_INVALIDDATA;
281 return AVERROR_INVALIDDATA;
283 if (caf->bytes_per_packet > 0 && caf->frames_per_packet > 0) {
284 if (caf->data_size > 0)
285 st->nb_frames = (caf->data_size / caf->bytes_per_packet) * caf->frames_per_packet;
286 } else if (st->nb_index_entries) {
287 st->codec->bit_rate = st->codec->sample_rate * caf->data_size * 8 /
290 av_log(s, AV_LOG_ERROR, "Missing packet table. It is required when "
291 "block size or frame size are variable.\n");
292 return AVERROR_INVALIDDATA;
294 s->file_size = avio_size(pb);
295 s->file_size = FFMAX(0, s->file_size);
297 av_set_pts_info(st, 64, 1, st->codec->sample_rate);
300 /* position the stream at the start of data */
301 if (caf->data_size >= 0)
302 avio_seek(pb, caf->data_start, SEEK_SET);
307 #define CAF_MAX_PKT_SIZE 4096
309 static int read_packet(AVFormatContext *s, AVPacket *pkt)
311 AVIOContext *pb = s->pb;
312 AVStream *st = s->streams[0];
313 CaffContext *caf = s->priv_data;
314 int res, pkt_size = 0, pkt_frames = 0;
315 int64_t left = CAF_MAX_PKT_SIZE;
320 /* don't read past end of data chunk */
321 if (caf->data_size > 0) {
322 left = (caf->data_start + caf->data_size) - avio_tell(pb);
327 pkt_frames = caf->frames_per_packet;
328 pkt_size = caf->bytes_per_packet;
330 if (pkt_size > 0 && pkt_frames == 1) {
331 pkt_size = (CAF_MAX_PKT_SIZE / pkt_size) * pkt_size;
332 pkt_size = FFMIN(pkt_size, left);
333 pkt_frames = pkt_size / caf->bytes_per_packet;
334 } else if (st->nb_index_entries) {
335 if (caf->packet_cnt < st->nb_index_entries - 1) {
336 pkt_size = st->index_entries[caf->packet_cnt + 1].pos - st->index_entries[caf->packet_cnt].pos;
337 pkt_frames = st->index_entries[caf->packet_cnt + 1].timestamp - st->index_entries[caf->packet_cnt].timestamp;
338 } else if (caf->packet_cnt == st->nb_index_entries - 1) {
339 pkt_size = caf->num_bytes - st->index_entries[caf->packet_cnt].pos;
340 pkt_frames = st->duration - st->index_entries[caf->packet_cnt].timestamp;
346 if (pkt_size == 0 || pkt_frames == 0 || pkt_size > left)
349 res = av_get_packet(pb, pkt, pkt_size);
354 pkt->stream_index = 0;
355 pkt->dts = pkt->pts = caf->frame_cnt;
358 caf->frame_cnt += pkt_frames;
363 static int read_seek(AVFormatContext *s, int stream_index,
364 int64_t timestamp, int flags)
366 AVStream *st = s->streams[0];
367 CaffContext *caf = s->priv_data;
370 timestamp = FFMAX(timestamp, 0);
372 if (caf->frames_per_packet > 0 && caf->bytes_per_packet > 0) {
373 /* calculate new byte position based on target frame position */
374 pos = caf->bytes_per_packet * timestamp / caf->frames_per_packet;
375 if (caf->data_size > 0)
376 pos = FFMIN(pos, caf->data_size);
377 caf->packet_cnt = pos / caf->bytes_per_packet;
378 caf->frame_cnt = caf->frames_per_packet * caf->packet_cnt;
379 } else if (st->nb_index_entries) {
380 caf->packet_cnt = av_index_search_timestamp(st, timestamp, flags);
381 caf->frame_cnt = st->index_entries[caf->packet_cnt].timestamp;
382 pos = st->index_entries[caf->packet_cnt].pos;
387 avio_seek(s->pb, pos + caf->data_start, SEEK_SET);
391 AVInputFormat ff_caf_demuxer = {
393 NULL_IF_CONFIG_SMALL("Apple Core Audio Format"),
400 .codec_tag = (const AVCodecTag*[]){ff_codec_caf_tags, 0},