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"
35 int bytes_per_packet; ///< bytes in a packet, or 0 if variable
36 int frames_per_packet; ///< frames in a packet, or 0 if variable
37 int64_t num_bytes; ///< total number of bytes in stream
39 int64_t packet_cnt; ///< packet counter
40 int64_t frame_cnt; ///< frame counter
42 int64_t data_start; ///< data start position, in bytes
43 int64_t data_size; ///< raw data size, in bytes
46 static int probe(AVProbeData *p)
48 if (AV_RB32(p->buf) == MKBETAG('c','a','f','f') && AV_RB16(&p->buf[4]) == 1)
49 return AVPROBE_SCORE_MAX;
53 /** Read audio description chunk */
54 static int read_desc_chunk(AVFormatContext *s)
56 ByteIOContext *pb = s->pb;
57 CaffContext *caf = s->priv_data;
61 /* new audio stream */
62 st = av_new_stream(s, 0);
64 return AVERROR(ENOMEM);
66 /* parse format description */
67 st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
68 st->codec->sample_rate = av_int2dbl(get_be64(pb));
69 st->codec->codec_tag = get_be32(pb);
71 caf->bytes_per_packet = get_be32(pb);
72 st->codec->block_align = caf->bytes_per_packet;
73 caf->frames_per_packet = get_be32(pb);
74 st->codec->channels = get_be32(pb);
75 st->codec->bits_per_coded_sample = get_be32(pb);
77 /* calculate bit rate for constant size packets */
78 if (caf->frames_per_packet > 0 && caf->bytes_per_packet > 0) {
79 st->codec->bit_rate = (uint64_t)st->codec->sample_rate * (uint64_t)caf->bytes_per_packet * 8
80 / (uint64_t)caf->frames_per_packet;
82 st->codec->bit_rate = 0;
86 if (st->codec->codec_tag == MKBETAG('l','p','c','m'))
87 st->codec->codec_id = ff_mov_get_lpcm_codec_id(st->codec->bits_per_coded_sample, (flags ^ 0x2) | 0x4);
89 st->codec->codec_id = ff_codec_get_id(ff_codec_caf_tags, st->codec->codec_tag);
93 /** Read magic cookie chunk */
94 static int read_kuki_chunk(AVFormatContext *s, int64_t size)
96 ByteIOContext *pb = s->pb;
97 AVStream *st = s->streams[0];
99 if (size < 0 || size > INT_MAX - FF_INPUT_BUFFER_PADDING_SIZE)
102 if (st->codec->codec_id == CODEC_ID_AAC) {
103 /* The magic cookie format for AAC is an mp4 esds atom.
104 The lavc AAC decoder requires the data from the codec specific
105 description as extradata input. */
109 strt = url_ftell(pb);
110 ff_mov_read_esds(s, pb, atom);
111 skip = size - (url_ftell(pb) - strt);
112 if (skip < 0 || !st->codec->extradata ||
113 st->codec->codec_id != CODEC_ID_AAC) {
114 av_log(s, AV_LOG_ERROR, "invalid AAC magic cookie\n");
115 return AVERROR_INVALIDDATA;
118 } else if (st->codec->codec_id == CODEC_ID_ALAC) {
119 #define ALAC_PREAMBLE 12
120 #define ALAC_HEADER 36
121 if (size < ALAC_PREAMBLE + ALAC_HEADER) {
122 av_log(s, AV_LOG_ERROR, "invalid ALAC magic cookie\n");
124 return AVERROR_INVALIDDATA;
126 url_fskip(pb, ALAC_PREAMBLE);
127 st->codec->extradata = av_mallocz(ALAC_HEADER + FF_INPUT_BUFFER_PADDING_SIZE);
128 if (!st->codec->extradata)
129 return AVERROR(ENOMEM);
130 get_buffer(pb, st->codec->extradata, ALAC_HEADER);
131 st->codec->extradata_size = ALAC_HEADER;
132 url_fskip(pb, size - ALAC_PREAMBLE - ALAC_HEADER);
134 st->codec->extradata = av_mallocz(size + FF_INPUT_BUFFER_PADDING_SIZE);
135 if (!st->codec->extradata)
136 return AVERROR(ENOMEM);
137 get_buffer(pb, st->codec->extradata, size);
138 st->codec->extradata_size = size;
144 /** Read packet table chunk */
145 static int read_pakt_chunk(AVFormatContext *s, int64_t size)
147 ByteIOContext *pb = s->pb;
148 AVStream *st = s->streams[0];
149 CaffContext *caf = s->priv_data;
150 int64_t pos = 0, ccount;
153 ccount = url_ftell(pb);
155 num_packets = get_be64(pb);
156 if (num_packets < 0 || INT32_MAX / sizeof(AVIndexEntry) < num_packets)
157 return AVERROR_INVALIDDATA;
159 st->nb_frames = get_be64(pb); /* valid frames */
160 st->nb_frames += get_be32(pb); /* priming frames */
161 st->nb_frames += get_be32(pb); /* remainder frames */
164 for (i = 0; i < num_packets; i++) {
165 av_add_index_entry(s->streams[0], pos, st->duration, 0, 0, AVINDEX_KEYFRAME);
166 pos += caf->bytes_per_packet ? caf->bytes_per_packet : ff_mp4_read_descr_len(pb);
167 st->duration += caf->frames_per_packet ? caf->frames_per_packet : ff_mp4_read_descr_len(pb);
170 if (url_ftell(pb) - ccount != size) {
171 av_log(s, AV_LOG_ERROR, "error reading packet table\n");
175 caf->num_bytes = pos;
179 /** Read information chunk */
180 static void read_info_chunk(AVFormatContext *s, int64_t size)
182 ByteIOContext *pb = s->pb;
184 unsigned int nb_entries = get_be32(pb);
185 for (i = 0; i < nb_entries; i++) {
188 get_strz(pb, key, sizeof(key));
189 get_strz(pb, value, sizeof(value));
190 av_metadata_set2(&s->metadata, key, value, 0);
194 static int read_header(AVFormatContext *s,
195 AVFormatParameters *ap)
197 ByteIOContext *pb = s->pb;
198 CaffContext *caf = s->priv_data;
204 url_fskip(pb, 8); /* magic, version, file flags */
206 /* audio description chunk */
207 if (get_be32(pb) != MKBETAG('d','e','s','c')) {
208 av_log(s, AV_LOG_ERROR, "desc chunk not present\n");
209 return AVERROR_INVALIDDATA;
213 return AVERROR_INVALIDDATA;
215 ret = read_desc_chunk(s);
220 /* parse each chunk */
222 while (!url_feof(pb)) {
224 /* stop at data chunk if seeking is not supported or
225 data chunk size is unknown */
226 if (found_data && (caf->data_size < 0 || url_is_streamed(pb)))
235 case MKBETAG('d','a','t','a'):
236 url_fskip(pb, 4); /* edit count */
237 caf->data_start = url_ftell(pb);
238 caf->data_size = size < 0 ? -1 : size - 4;
239 if (caf->data_size > 0 && !url_is_streamed(pb))
240 url_fskip(pb, caf->data_size);
244 /* magic cookie chunk */
245 case MKBETAG('k','u','k','i'):
246 if (read_kuki_chunk(s, size))
247 return AVERROR_INVALIDDATA;
250 /* packet table chunk */
251 case MKBETAG('p','a','k','t'):
252 if (read_pakt_chunk(s, size))
253 return AVERROR_INVALIDDATA;
256 case MKBETAG('i','n','f','o'):
257 read_info_chunk(s, size);
261 #define _(x) ((x) >= ' ' ? (x) : ' ')
262 av_log(s, AV_LOG_WARNING, "skipping CAF chunk: %08X (%c%c%c%c)\n",
263 tag, _(tag>>24), _((tag>>16)&0xFF), _((tag>>8)&0xFF), _(tag&0xFF));
265 case MKBETAG('f','r','e','e'):
267 return AVERROR_INVALIDDATA;
274 return AVERROR_INVALIDDATA;
276 if (caf->bytes_per_packet > 0 && caf->frames_per_packet > 0) {
277 if (caf->data_size > 0)
278 st->nb_frames = (caf->data_size / caf->bytes_per_packet) * caf->frames_per_packet;
279 } else if (st->nb_index_entries) {
280 st->codec->bit_rate = st->codec->sample_rate * caf->data_size * 8 /
283 av_log(s, AV_LOG_ERROR, "Missing packet table. It is required when "
284 "block size or frame size are variable.\n");
285 return AVERROR_INVALIDDATA;
287 s->file_size = url_fsize(pb);
288 s->file_size = FFMAX(0, s->file_size);
290 av_set_pts_info(st, 64, 1, st->codec->sample_rate);
293 /* position the stream at the start of data */
294 if (caf->data_size >= 0)
295 url_fseek(pb, caf->data_start, SEEK_SET);
300 #define CAF_MAX_PKT_SIZE 4096
302 static int read_packet(AVFormatContext *s, AVPacket *pkt)
304 ByteIOContext *pb = s->pb;
305 AVStream *st = s->streams[0];
306 CaffContext *caf = s->priv_data;
307 int res, pkt_size = 0, pkt_frames = 0;
308 int64_t left = CAF_MAX_PKT_SIZE;
313 /* don't read past end of data chunk */
314 if (caf->data_size > 0) {
315 left = (caf->data_start + caf->data_size) - url_ftell(pb);
320 pkt_frames = caf->frames_per_packet;
321 pkt_size = caf->bytes_per_packet;
323 if (pkt_size > 0 && pkt_frames == 1) {
324 pkt_size = (CAF_MAX_PKT_SIZE / pkt_size) * pkt_size;
325 pkt_size = FFMIN(pkt_size, left);
326 pkt_frames = pkt_size / caf->bytes_per_packet;
327 } else if (st->nb_index_entries) {
328 if (caf->packet_cnt < st->nb_index_entries - 1) {
329 pkt_size = st->index_entries[caf->packet_cnt + 1].pos - st->index_entries[caf->packet_cnt].pos;
330 pkt_frames = st->index_entries[caf->packet_cnt + 1].timestamp - st->index_entries[caf->packet_cnt].timestamp;
331 } else if (caf->packet_cnt == st->nb_index_entries - 1) {
332 pkt_size = caf->num_bytes - st->index_entries[caf->packet_cnt].pos;
333 pkt_frames = st->duration - st->index_entries[caf->packet_cnt].timestamp;
339 if (pkt_size == 0 || pkt_frames == 0 || pkt_size > left)
342 res = av_get_packet(pb, pkt, pkt_size);
347 pkt->stream_index = 0;
348 pkt->dts = pkt->pts = caf->frame_cnt;
351 caf->frame_cnt += pkt_frames;
356 static int read_seek(AVFormatContext *s, int stream_index,
357 int64_t timestamp, int flags)
359 AVStream *st = s->streams[0];
360 CaffContext *caf = s->priv_data;
363 timestamp = FFMAX(timestamp, 0);
365 if (caf->frames_per_packet > 0 && caf->bytes_per_packet > 0) {
366 /* calculate new byte position based on target frame position */
367 pos = caf->bytes_per_packet * timestamp / caf->frames_per_packet;
368 if (caf->data_size > 0)
369 pos = FFMIN(pos, caf->data_size);
370 caf->packet_cnt = pos / caf->bytes_per_packet;
371 caf->frame_cnt = caf->frames_per_packet * caf->packet_cnt;
372 } else if (st->nb_index_entries) {
373 caf->packet_cnt = av_index_search_timestamp(st, timestamp, flags);
374 caf->frame_cnt = st->index_entries[caf->packet_cnt].timestamp;
375 pos = st->index_entries[caf->packet_cnt].pos;
380 url_fseek(s->pb, pos + caf->data_start, SEEK_SET);
384 AVInputFormat ff_caf_demuxer = {
386 NULL_IF_CONFIG_SMALL("Apple Core Audio Format"),
393 .codec_tag = (const AVCodecTag*[]){ff_codec_caf_tags, 0},