isom.h 9.8 KB
Newer Older
1 2
/*
 * ISO Media common code
3
 * copyright (c) 2001 Fabrice Bellard
4 5 6
 * copyright (c) 2002 Francois Revol <revol@free.fr>
 * copyright (c) 2006 Baptiste Coudurier <baptiste.coudurier@free.fr>
 *
7 8 9
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
10 11
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
12
 * version 2.1 of the License, or (at your option) any later version.
13
 *
14
 * FFmpeg is distributed in the hope that it will be useful,
15 16 17 18 19
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
20
 * License along with FFmpeg; if not, write to the Free Software
21 22 23
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

24 25
#ifndef AVFORMAT_ISOM_H
#define AVFORMAT_ISOM_H
26

27
#include "avio.h"
28
#include "internal.h"
29
#include "dv.h"
30

31
/* isom.c */
32
extern const AVCodecTag ff_mp4_obj_type[];
33 34
extern const AVCodecTag ff_codec_movvideo_tags[];
extern const AVCodecTag ff_codec_movaudio_tags[];
35
extern const AVCodecTag ff_codec_movsubtitle_tags[];
36

37 38
int ff_mov_iso639_to_lang(const char lang[4], int mp4);
int ff_mov_lang_to_iso639(unsigned code, char to[4]);
39

40 41
struct AVAESCTR;

42 43 44 45 46
/* the QuickTime file format is quite convoluted...
 * it has lots of index tables, each indexing something in another one...
 * Here we just use what is needed to read the chunks
 */

47
typedef struct MOVStts {
48 49
    int count;
    int duration;
50
} MOVStts;
51

52
typedef struct MOVStsc {
53 54 55 56 57
    int first;
    int count;
    int id;
} MOVStsc;

58 59 60 61 62 63
typedef struct MOVElst {
    int64_t duration;
    int64_t time;
    float rate;
} MOVElst;

64
typedef struct MOVDref {
65 66
    uint32_t type;
    char *path;
67 68 69 70
    char *dir;
    char volume[28];
    char filename[64];
    int16_t nlvl_to, nlvl_from;
71 72
} MOVDref;

73
typedef struct MOVAtom {
74 75 76 77 78 79
    uint32_t type;
    int64_t size; /* total size (excluding the size and type fields) */
} MOVAtom;

struct MOVParseTableEntry;

80
typedef struct MOVFragment {
81 82 83
    unsigned track_id;
    uint64_t base_data_offset;
    uint64_t moof_offset;
84
    uint64_t implicit_offset;
85 86 87 88
    unsigned stsd_id;
    unsigned duration;
    unsigned size;
    unsigned flags;
89
    int64_t time;
90 91
} MOVFragment;

92
typedef struct MOVTrackExt {
93 94 95 96 97 98 99
    unsigned track_id;
    unsigned stsd_id;
    unsigned duration;
    unsigned size;
    unsigned flags;
} MOVTrackExt;

100
typedef struct MOVSbgp {
101 102 103 104
    unsigned int count;
    unsigned int index;
} MOVSbgp;

105 106 107
typedef struct MOVFragmentIndexItem {
    int64_t moof_offset;
    int64_t time;
108
    int headers_read;
109 110 111 112 113 114 115 116 117
} MOVFragmentIndexItem;

typedef struct MOVFragmentIndex {
    unsigned track_id;
    unsigned item_count;
    unsigned current_item;
    MOVFragmentIndexItem *items;
} MOVFragmentIndex;

118
typedef struct MOVStreamContext {
119
    AVIOContext *pb;
120
    int pb_is_copied;
121
    int ffindex;          ///< AVStream index
122 123 124 125 126 127 128 129 130
    int next_chunk;
    unsigned int chunk_count;
    int64_t *chunk_offsets;
    unsigned int stts_count;
    MOVStts *stts_data;
    unsigned int ctts_count;
    MOVStts *ctts_data;
    unsigned int stsc_count;
    MOVStsc *stsc_data;
131 132
    int stsc_index;
    int stsc_sample;
133 134
    unsigned int stps_count;
    unsigned *stps_data;  ///< partial sync sample for mpeg-2 open gop
135 136
    MOVElst *elst_data;
    unsigned int elst_count;
137 138
    int ctts_index;
    int ctts_sample;
139
    unsigned int sample_size; ///< may contain value calculated from stsd or value from stsz atom
140
    unsigned int stsz_sample_size; ///< always contains sample size from stsz atom
141 142
    unsigned int sample_count;
    int *sample_sizes;
143
    int keyframe_absent;
144 145 146
    unsigned int keyframe_count;
    int *keyframes;
    int time_scale;
147
    int64_t time_offset;  ///< time offset of the edit list entries
148 149 150 151 152
    int current_sample;
    unsigned int bytes_per_frame;
    unsigned int samples_per_frame;
    int dv_audio_container;
    int pseudo_stream_id; ///< -1 means demux all ids
153
    int16_t audio_cid;    ///< stsd audio compression id
154 155 156
    unsigned drefs_count;
    MOVDref *drefs;
    int dref_id;
157
    int timecode_track;
158 159 160
    int width;            ///< tkhd width
    int height;           ///< tkhd height
    int dts_shift;        ///< dts shift when ctts is negative
161 162
    uint32_t palette[256];
    int has_palette;
163
    int64_t data_size;
164
    uint32_t tmcd_flags;  ///< tmcd track flags
165
    int64_t track_end;    ///< used for dts generation in fragmented movie files
166
    int start_pad;        ///< amount of samples to skip due to enc-dec delay
167 168
    unsigned int rap_group_count;
    MOVSbgp *rap_group;
169 170 171

    int nb_frames_for_fps;
    int64_t duration_for_fps;
172

173 174 175 176 177 178
    /** extradata array (and size) for multiple stsd */
    uint8_t **extradata;
    int *extradata_size;
    int last_stsd_index;
    int stsd_count;

179
    int32_t *display_matrix;
180 181
    uint32_t format;

182
    int has_sidx;  // If there is an sidx entry for this stream.
183 184 185 186 187
    struct {
        int use_subsamples;
        uint8_t* auxiliary_info;
        uint8_t* auxiliary_info_end;
        uint8_t* auxiliary_info_pos;
188 189 190
        uint8_t auxiliary_info_default_size;
        uint8_t* auxiliary_info_sizes;
        size_t auxiliary_info_sizes_count;
191 192
        struct AVAESCTR* aes_ctr;
    } cenc;
193 194 195
} MOVStreamContext;

typedef struct MOVContext {
196
    const AVClass *class; ///< class for private options
197 198
    AVFormatContext *fc;
    int time_scale;
199 200 201
    int64_t duration;     ///< duration of the longest track
    int found_moov;       ///< 'moov' atom has been found
    int found_mdat;       ///< 'mdat' atom has been found
202
    int found_hdlr_mdta;  ///< 'hdlr' atom with type 'mdta' has been found
203
    int trak_index;       ///< Index of the current 'trak'
204 205
    char **meta_keys;
    unsigned meta_keys_count;
206 207
    DVDemuxContext *dv_demux;
    AVFormatContext *dv_fctx;
208
    int isom;             ///< 1 if file is ISO Media (mp4/3gp)
209 210 211
    MOVFragment fragment; ///< current fragment in moof atom
    MOVTrackExt *trex_data;
    unsigned trex_count;
212
    int itunes_metadata;  ///< metadata are itunes style
213
    int handbrake_version;
214 215
    int *chapter_tracks;
    unsigned int nb_chapter_tracks;
216
    int use_absolute_path;
217
    int ignore_editlist;
218
    int ignore_chapters;
219
    int seek_individually;
220
    int64_t next_root_atom; ///< offset of the next root atom
221
    int export_all;
222
    int export_xmp;
223 224
    int *bitrates;          ///< bitrates read before streams creation
    int bitrates_count;
225
    int moov_retry;
226 227 228 229
    int use_mfra_for;
    int has_looked_for_mfra;
    MOVFragmentIndex** fragment_index_data;
    unsigned fragment_index_count;
230
    int fragment_index_complete;
231
    int atom_depth;
232 233 234 235 236 237 238 239
    unsigned int aax_mode;  ///< 'aax' file has been detected
    uint8_t file_key[20];
    uint8_t file_iv[20];
    void *activation_bytes;
    int activation_bytes_size;
    void *audible_fixed_key;
    int audible_fixed_key_size;
    struct AVAES *aes_decrypt;
240 241
    uint8_t *decryption_key;
    int decryption_key_len;
242
    int enable_drefs;
243 244
} MOVContext;

245 246 247
int ff_mp4_read_descr_len(AVIOContext *pb);
int ff_mp4_read_descr(AVFormatContext *fc, AVIOContext *pb, int *tag);
int ff_mp4_read_dec_config_descr(AVFormatContext *fc, AVStream *st, AVIOContext *pb);
248
void ff_mp4_parse_es_descr(AVIOContext *pb, int *es_id);
249

Alex Converse's avatar
Alex Converse committed
250
#define MP4ODescrTag                    0x01
251 252 253 254
#define MP4IODescrTag                   0x02
#define MP4ESDescrTag                   0x03
#define MP4DecConfigDescrTag            0x04
#define MP4DecSpecificDescrTag          0x05
Alex Converse's avatar
Alex Converse committed
255
#define MP4SLDescrTag                   0x06
256

257 258 259 260 261 262
#define MOV_TFHD_BASE_DATA_OFFSET       0x01
#define MOV_TFHD_STSD_ID                0x02
#define MOV_TFHD_DEFAULT_DURATION       0x08
#define MOV_TFHD_DEFAULT_SIZE           0x10
#define MOV_TFHD_DEFAULT_FLAGS          0x20
#define MOV_TFHD_DURATION_IS_EMPTY  0x010000
263
#define MOV_TFHD_DEFAULT_BASE_IS_MOOF 0x020000
264

265 266 267 268 269 270 271
#define MOV_TRUN_DATA_OFFSET            0x01
#define MOV_TRUN_FIRST_SAMPLE_FLAGS     0x04
#define MOV_TRUN_SAMPLE_DURATION       0x100
#define MOV_TRUN_SAMPLE_SIZE           0x200
#define MOV_TRUN_SAMPLE_FLAGS          0x400
#define MOV_TRUN_SAMPLE_CTS            0x800

272 273 274 275 276 277 278 279 280 281
#define MOV_FRAG_SAMPLE_FLAG_DEGRADATION_PRIORITY_MASK 0x0000ffff
#define MOV_FRAG_SAMPLE_FLAG_IS_NON_SYNC               0x00010000
#define MOV_FRAG_SAMPLE_FLAG_PADDING_MASK              0x000e0000
#define MOV_FRAG_SAMPLE_FLAG_REDUNDANCY_MASK           0x00300000
#define MOV_FRAG_SAMPLE_FLAG_DEPENDED_MASK             0x00c00000
#define MOV_FRAG_SAMPLE_FLAG_DEPENDS_MASK              0x03000000

#define MOV_FRAG_SAMPLE_FLAG_DEPENDS_NO                0x02000000
#define MOV_FRAG_SAMPLE_FLAG_DEPENDS_YES               0x01000000

282 283 284 285 286
#define MOV_TKHD_FLAG_ENABLED       0x0001
#define MOV_TKHD_FLAG_IN_MOVIE      0x0002
#define MOV_TKHD_FLAG_IN_PREVIEW    0x0004
#define MOV_TKHD_FLAG_IN_POSTER     0x0008

287 288 289 290 291 292 293 294 295 296 297 298 299
#define TAG_IS_AVCI(tag)                    \
    ((tag) == MKTAG('a', 'i', '5', 'p') ||  \
     (tag) == MKTAG('a', 'i', '5', 'q') ||  \
     (tag) == MKTAG('a', 'i', '5', '2') ||  \
     (tag) == MKTAG('a', 'i', '5', '3') ||  \
     (tag) == MKTAG('a', 'i', '5', '5') ||  \
     (tag) == MKTAG('a', 'i', '5', '6') ||  \
     (tag) == MKTAG('a', 'i', '1', 'p') ||  \
     (tag) == MKTAG('a', 'i', '1', 'q') ||  \
     (tag) == MKTAG('a', 'i', '1', '2') ||  \
     (tag) == MKTAG('a', 'i', '1', '3') ||  \
     (tag) == MKTAG('a', 'i', '1', '5') ||  \
     (tag) == MKTAG('a', 'i', '1', '6') ||  \
300
     (tag) == MKTAG('a', 'i', 'v', 'x') ||  \
301 302 303
     (tag) == MKTAG('A', 'V', 'i', 'n'))


304
int ff_mov_read_esds(AVFormatContext *fc, AVIOContext *pb);
305
enum AVCodecID ff_mov_get_lpcm_codec_id(int bps, int flags);
306

307
int ff_mov_read_stsd_entries(MOVContext *c, AVIOContext *pb, int entries);
308
void ff_mov_write_chan(AVIOContext *pb, int64_t channel_layout);
309

310
#define FF_MOV_FLAG_MFRA_AUTO -1
311 312 313
#define FF_MOV_FLAG_MFRA_DTS 1
#define FF_MOV_FLAG_MFRA_PTS 2

314
#endif /* AVFORMAT_ISOM_H */