/*
* MOV demuxer
- * Copyright (c) 2001 Fabrice Bellard.
+ * Copyright (c) 2001 Fabrice Bellard
+ * Copyright (c) 2009 Baptiste Coudurier <baptiste dot coudurier at gmail dot com>
*
* This file is part of FFmpeg.
*
#include <limits.h>
//#define DEBUG
+//#define DEBUG_METADATA
+//#define MOV_EXPORT_ALL_METADATA
+#include "libavutil/intreadwrite.h"
+#include "libavutil/avstring.h"
#include "avformat.h"
#include "riff.h"
#include "isom.h"
-#include "dv.h"
#include "libavcodec/mpeg4audio.h"
#include "libavcodec/mpegaudiodata.h"
+#include "libavcodec/get_bits.h"
-#ifdef CONFIG_ZLIB
+#if CONFIG_ZLIB
#include <zlib.h>
#endif
#undef NDEBUG
#include <assert.h>
-/* 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
- */
-
-typedef struct {
- int first;
- int count;
- int id;
-} MOVStsc;
-
-typedef struct {
- uint32_t type;
- char *path;
-} MOVDref;
-
-typedef struct {
- uint32_t type;
- int64_t offset;
- int64_t size; /* total size (excluding the size and type fields) */
-} MOVAtom;
-
-struct MOVParseTableEntry;
-
-typedef struct {
- unsigned track_id;
- uint64_t base_data_offset;
- uint64_t moof_offset;
- unsigned stsd_id;
- unsigned duration;
- unsigned size;
- unsigned flags;
-} MOVFragment;
-
-typedef struct {
- unsigned track_id;
- unsigned stsd_id;
- unsigned duration;
- unsigned size;
- unsigned flags;
-} MOVTrackExt;
-
-typedef struct MOVStreamContext {
- ByteIOContext *pb;
- int ffindex; /* the ffmpeg stream id */
- 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 edit_count; /* number of 'edit' (elst atom) */
- unsigned int sample_to_chunk_sz;
- MOVStsc *sample_to_chunk;
- int sample_to_ctime_index;
- int sample_to_ctime_sample;
- unsigned int sample_size;
- unsigned int sample_count;
- int *sample_sizes;
- unsigned int keyframe_count;
- int *keyframes;
- int time_scale;
- int time_rate;
- 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
- int16_t audio_cid; ///< stsd audio compression id
- unsigned drefs_count;
- MOVDref *drefs;
- int dref_id;
- int wrong_dts; ///< dts are wrong due to negative ctts
-} MOVStreamContext;
-
-typedef struct MOVContext {
- AVFormatContext *fc;
- int time_scale;
- int64_t duration; /* duration of the longest track */
- int found_moov; /* when both 'moov' and 'mdat' sections has been found */
- int found_mdat; /* we suppose we have enough data to read the file */
- AVPaletteControl palette_control;
- DVDemuxContext *dv_demux;
- AVFormatContext *dv_fctx;
- int isom; /* 1 if file is ISO Media (mp4/3gp) */
- MOVFragment fragment; ///< current fragment in moof atom
- MOVTrackExt *trex_data;
- unsigned trex_count;
- int itunes_metadata; ///< metadata are itunes style
-} MOVContext;
-
-
/* XXX: it's the first time I make a recursive parser I think... sorry if it's ugly :P */
/* those functions parse an atom */
static const MOVParseTableEntry mov_default_parse_table[];
+static int mov_metadata_trkn(MOVContext *c, ByteIOContext *pb, unsigned len)
+{
+ char buf[16];
+
+ get_be16(pb); // unknown
+ snprintf(buf, sizeof(buf), "%d", get_be16(pb));
+ av_metadata_set(&c->fc->metadata, "track", buf);
+
+ get_be16(pb); // total tracks
+
+ return 0;
+}
+
+static int mov_read_udta_string(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
+{
+#ifdef MOV_EXPORT_ALL_METADATA
+ char tmp_key[5];
+#endif
+ char str[1024], key2[16], language[4] = {0};
+ const char *key = NULL;
+ uint16_t str_size;
+ int (*parse)(MOVContext*, ByteIOContext*, unsigned) = NULL;
+
+ switch (atom.type) {
+ case MKTAG(0xa9,'n','a','m'): key = "title"; break;
+ case MKTAG(0xa9,'a','u','t'):
+ case MKTAG(0xa9,'A','R','T'):
+ case MKTAG(0xa9,'w','r','t'): key = "author"; break;
+ case MKTAG(0xa9,'c','p','y'): key = "copyright"; break;
+ case MKTAG(0xa9,'c','m','t'):
+ case MKTAG(0xa9,'i','n','f'): key = "comment"; break;
+ case MKTAG(0xa9,'a','l','b'): key = "album"; break;
+ case MKTAG(0xa9,'d','a','y'): key = "year"; break;
+ case MKTAG(0xa9,'g','e','n'): key = "genre"; break;
+ case MKTAG(0xa9,'t','o','o'):
+ case MKTAG(0xa9,'e','n','c'): key = "muxer"; break;
+ case MKTAG( 't','r','k','n'): key = "track";
+ parse = mov_metadata_trkn; break;
+ }
+
+ if (c->itunes_metadata && atom.size > 8) {
+ int data_size = get_be32(pb);
+ int tag = get_le32(pb);
+ if (tag == MKTAG('d','a','t','a')) {
+ get_be32(pb); // type
+ get_be32(pb); // unknown
+ str_size = data_size - 16;
+ atom.size -= 16;
+ } else return 0;
+ } else if (atom.size > 4 && key && !c->itunes_metadata) {
+ str_size = get_be16(pb); // string length
+ ff_mov_lang_to_iso639(get_be16(pb), language);
+ atom.size -= 4;
+ } else
+ str_size = atom.size;
+
+#ifdef MOV_EXPORT_ALL_METADATA
+ if (!key) {
+ snprintf(tmp_key, 5, "%.4s", (char*)&atom.type);
+ key = tmp_key;
+ }
+#endif
+
+ if (!key)
+ return 0;
+ if (atom.size < 0)
+ return -1;
+
+ str_size = FFMIN3(sizeof(str)-1, str_size, atom.size);
+
+ if (parse)
+ parse(c, pb, str_size);
+ else {
+ get_buffer(pb, str, str_size);
+ str[str_size] = 0;
+ av_metadata_set(&c->fc->metadata, key, str);
+ if (*language && strcmp(language, "und")) {
+ snprintf(key2, sizeof(key2), "%s-%s", key, language);
+ av_metadata_set(&c->fc->metadata, key2, str);
+ }
+ }
+#ifdef DEBUG_METADATA
+ av_log(c->fc, AV_LOG_DEBUG, "lang \"%3s\" ", language);
+ av_log(c->fc, AV_LOG_DEBUG, "tag \"%s\" value \"%s\" atom \"%.4s\" %d %lld\n",
+ key, str, (char*)&atom.type, str_size, atom.size);
+#endif
+
+ return 0;
+}
+
static int mov_read_default(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
{
int64_t total_size = 0;
if (atom.size < 0)
atom.size = INT64_MAX;
while(((total_size + 8) < atom.size) && !url_feof(pb) && !err) {
+ int (*parse)(MOVContext*, ByteIOContext*, MOVAtom) = NULL;
a.size = atom.size;
a.type=0;
if(atom.size >= 8) {
break;
a.size = FFMIN(a.size, atom.size - total_size);
- for (i = 0; mov_default_parse_table[i].type != 0
- && mov_default_parse_table[i].type != a.type; i++)
- /* empty */;
+ for (i = 0; mov_default_parse_table[i].type; i++)
+ if (mov_default_parse_table[i].type == a.type) {
+ parse = mov_default_parse_table[i].parse;
+ break;
+ }
+
+ // container is user data
+ if (!parse && (atom.type == MKTAG('u','d','t','a') ||
+ atom.type == MKTAG('i','l','s','t')))
+ parse = mov_read_udta_string;
- if (mov_default_parse_table[i].type == 0) { /* skip leaf atoms data */
+ if (!parse) { /* skip leaf atoms data */
url_fskip(pb, a.size);
} else {
int64_t start_pos = url_ftell(pb);
int64_t left;
- err = mov_default_parse_table[i].parse(c, pb, a);
+ err = parse(c, pb, a);
if (url_is_streamed(pb) && c->found_moov && c->found_mdat)
break;
left = a.size - url_ftell(pb) + start_pos;
entries = get_be32(pb);
if (entries >= UINT_MAX / sizeof(*sc->drefs))
return -1;
- sc->drefs_count = entries;
sc->drefs = av_mallocz(entries * sizeof(*sc->drefs));
+ if (!sc->drefs)
+ return AVERROR(ENOMEM);
+ sc->drefs_count = entries;
for (i = 0; i < sc->drefs_count; i++) {
MOVDref *dref = &sc->drefs[i];
static int mov_read_hdlr(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
{
- AVStream *st = c->fc->streams[c->fc->nb_streams-1];
+ AVStream *st;
uint32_t type;
uint32_t ctype;
+ if (c->fc->nb_streams < 1) // meta before first trak
+ return 0;
+
+ st = c->fc->streams[c->fc->nb_streams-1];
+
get_byte(pb); /* version */
get_be24(pb); /* flags */
ctype = get_le32(pb);
type = get_le32(pb); /* component subtype */
- dprintf(c->fc, "ctype= %c%c%c%c (0x%08x)\n", *((char *)&ctype), ((char *)&ctype)[1],
- ((char *)&ctype)[2], ((char *)&ctype)[3], (int) ctype);
- dprintf(c->fc, "stype= %c%c%c%c\n",
- *((char *)&type), ((char *)&type)[1], ((char *)&type)[2], ((char *)&type)[3]);
- if(!ctype)
- c->isom = 1;
+ dprintf(c->fc, "ctype= %.4s (0x%08x)\n", (char*)&ctype, ctype);
+ dprintf(c->fc, "stype= %.4s\n", (char*)&type);
+
if (type == MKTAG('v','i','d','e'))
st->codec->codec_type = CODEC_TYPE_VIDEO;
else if(type == MKTAG('s','o','u','n'))
st->codec->codec_type = CODEC_TYPE_AUDIO;
else if(type == MKTAG('m','1','a',' '))
st->codec->codec_id = CODEC_ID_MP2;
- else if(type == MKTAG('s','u','b','p')) {
+ else if(type == MKTAG('s','u','b','p'))
st->codec->codec_type = CODEC_TYPE_SUBTITLE;
- }
+
get_be32(pb); /* component manufacture */
get_be32(pb); /* component flags */
get_be32(pb); /* component flags mask */
#define MP4DecSpecificDescrTag 0x05
static const AVCodecTag mp4_audio_types[] = {
- { CODEC_ID_MP3ON4, 29 }, /* old mp3on4 draft */
- { CODEC_ID_MP3ON4, 32 }, /* layer 1 */
- { CODEC_ID_MP3ON4, 33 }, /* layer 2 */
- { CODEC_ID_MP3ON4, 34 }, /* layer 3 */
- { CODEC_ID_NONE, 0 },
+ { CODEC_ID_MP3ON4, AOT_PS }, /* old mp3on4 draft */
+ { CODEC_ID_MP3ON4, AOT_L1 }, /* layer 1 */
+ { CODEC_ID_MP3ON4, AOT_L2 }, /* layer 2 */
+ { CODEC_ID_MP3ON4, AOT_L3 }, /* layer 3 */
+ { CODEC_ID_MP4ALS, AOT_ALS }, /* MPEG-4 ALS */
+ { CODEC_ID_NONE, AOT_NULL },
};
static int mov_read_esds(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
const int den = get_be32(pb);
AVStream * const st = c->fc->streams[c->fc->nb_streams-1];
if (den != 0) {
- if ((st->sample_aspect_ratio.den && den != st->sample_aspect_ratio.den) ||
- (st->sample_aspect_ratio.num && num != st->sample_aspect_ratio.num))
+ if ((st->sample_aspect_ratio.den != 1 || st->sample_aspect_ratio.num) && // default
+ (den != st->sample_aspect_ratio.den || num != st->sample_aspect_ratio.num))
av_log(c->fc, AV_LOG_WARNING,
- "sample aspect ratio already set, overriding by 'pasp' atom\n");
+ "sample aspect ratio already set to %d:%d, overriding by 'pasp' atom\n",
+ st->sample_aspect_ratio.num, st->sample_aspect_ratio.den);
st->sample_aspect_ratio.num = num;
st->sample_aspect_ratio.den = den;
}
AVStream *st = c->fc->streams[c->fc->nb_streams-1];
MOVStreamContext *sc = st->priv_data;
int version = get_byte(pb);
- int lang;
+ char language[4] = {0};
+ unsigned lang;
if (version > 1)
return -1; /* unsupported */
st->duration = (version == 1) ? get_be64(pb) : get_be32(pb); /* duration */
lang = get_be16(pb); /* language */
- ff_mov_lang_to_iso639(lang, st->language);
+ if (ff_mov_lang_to_iso639(lang, language))
+ av_metadata_set(&st->metadata, "language", language);
get_be16(pb); /* quality */
return 0;
if(entries >= UINT_MAX/sizeof(int64_t))
return -1;
- sc->chunk_count = entries;
sc->chunk_offsets = av_malloc(entries * sizeof(int64_t));
if (!sc->chunk_offsets)
- return -1;
+ return AVERROR(ENOMEM);
+ sc->chunk_count = entries;
+
if (atom.type == MKTAG('s','t','c','o'))
for(i=0; i<entries; i++)
sc->chunk_offsets[i] = get_be32(pb);
unsigned int color_start, color_count, color_end;
unsigned char r, g, b;
+ st->codec->palctrl = av_malloc(sizeof(*st->codec->palctrl));
if (color_greyscale) {
int color_index, color_dec;
/* compute the greyscale palette */
color_dec = 256 / (color_count - 1);
for (j = 0; j < color_count; j++) {
r = g = b = color_index;
- c->palette_control.palette[j] =
+ st->codec->palctrl->palette[j] =
(r << 16) | (g << 8) | (b);
color_index -= color_dec;
if (color_index < 0)
color_table = ff_qt_default_palette_256;
for (j = 0; j < color_count; j++) {
- r = color_table[j * 4 + 0];
- g = color_table[j * 4 + 1];
- b = color_table[j * 4 + 2];
- c->palette_control.palette[j] =
+ r = color_table[j * 3 + 0];
+ g = color_table[j * 3 + 1];
+ b = color_table[j * 3 + 2];
+ st->codec->palctrl->palette[j] =
(r << 16) | (g << 8) | (b);
}
} else {
get_byte(pb);
b = get_byte(pb);
get_byte(pb);
- c->palette_control.palette[j] =
+ st->codec->palctrl->palette[j] =
(r << 16) | (g << 8) | (b);
}
}
}
- st->codec->palctrl = &c->palette_control;
st->codec->palctrl->palette_changed = 1;
- } else
- st->codec->palctrl = NULL;
+ }
} else if(st->codec->codec_type==CODEC_TYPE_AUDIO) {
int bits_per_sample, flags;
uint16_t version = get_be16(pb);
// ttxt stsd contains display flags, justification, background
// color, fonts, and default styles, so fake an atom to read it
MOVAtom fake_atom = { .size = size - (url_ftell(pb) - start_pos) };
- mov_read_glbl(c, pb, fake_atom);
+ if (format != AV_RL32("mp4s")) // mp4s contains a regular esds atom
+ mov_read_glbl(c, pb, fake_atom);
st->codec->codec_id= id;
+ st->codec->width = sc->width;
+ st->codec->height = sc->height;
} else {
/* other codec type, just skip (rtp, mp4s, tmcd ...) */
url_fskip(pb, size - (url_ftell(pb) - start_pos));
/* special codec parameters handling */
switch (st->codec->codec_id) {
-#ifdef CONFIG_DV_DEMUXER
+#if CONFIG_DV_DEMUXER
case CODEC_ID_DVAUDIO:
- c->dv_fctx = av_alloc_format_context();
+ c->dv_fctx = avformat_alloc_context();
c->dv_demux = dv_init_demux(c->dv_fctx);
if (!c->dv_demux) {
av_log(c->fc, AV_LOG_ERROR, "dv demux context init error\n");
#endif
/* no ifdef since parameters are always those */
case CODEC_ID_QCELP:
+ // force sample rate for qcelp when not stored in mov
+ if (st->codec->codec_tag != MKTAG('Q','c','l','p'))
+ st->codec->sample_rate = 8000;
st->codec->frame_size= 160;
st->codec->channels= 1; /* really needed */
break;
entries = get_be32(pb);
- if(entries >= UINT_MAX / sizeof(*sc->sample_to_chunk))
- return -1;
-
dprintf(c->fc, "track[%i].stsc.entries = %i\n", c->fc->nb_streams-1, entries);
- sc->sample_to_chunk_sz = entries;
- sc->sample_to_chunk = av_malloc(entries * sizeof(*sc->sample_to_chunk));
- if (!sc->sample_to_chunk)
+ if(entries >= UINT_MAX / sizeof(*sc->stsc_data))
return -1;
+ sc->stsc_data = av_malloc(entries * sizeof(*sc->stsc_data));
+ if (!sc->stsc_data)
+ return AVERROR(ENOMEM);
+ sc->stsc_count = entries;
+
for(i=0; i<entries; i++) {
- sc->sample_to_chunk[i].first = get_be32(pb);
- sc->sample_to_chunk[i].count = get_be32(pb);
- sc->sample_to_chunk[i].id = get_be32(pb);
+ sc->stsc_data[i].first = get_be32(pb);
+ sc->stsc_data[i].count = get_be32(pb);
+ sc->stsc_data[i].id = get_be32(pb);
}
return 0;
}
entries = get_be32(pb);
+ dprintf(c->fc, "keyframe_count = %d\n", entries);
+
if(entries >= UINT_MAX / sizeof(int))
return -1;
-
- sc->keyframe_count = entries;
-
- dprintf(c->fc, "keyframe_count = %d\n", sc->keyframe_count);
-
sc->keyframes = av_malloc(entries * sizeof(int));
if (!sc->keyframes)
- return -1;
+ return AVERROR(ENOMEM);
+ sc->keyframe_count = entries;
+
for(i=0; i<entries; i++) {
sc->keyframes[i] = get_be32(pb);
//dprintf(c->fc, "keyframes[]=%d\n", sc->keyframes[i]);
{
AVStream *st = c->fc->streams[c->fc->nb_streams-1];
MOVStreamContext *sc = st->priv_data;
- unsigned int i, entries, sample_size;
+ unsigned int i, entries, sample_size, field_size, num_bytes;
+ GetBitContext gb;
+ unsigned char* buf;
get_byte(pb); /* version */
get_be24(pb); /* flags */
- sample_size = get_be32(pb);
- if (!sc->sample_size) /* do not overwrite value computed in stsd */
- sc->sample_size = sample_size;
+ if (atom.type == MKTAG('s','t','s','z')) {
+ sample_size = get_be32(pb);
+ if (!sc->sample_size) /* do not overwrite value computed in stsd */
+ sc->sample_size = sample_size;
+ field_size = 32;
+ } else {
+ sample_size = 0;
+ get_be24(pb); /* reserved */
+ field_size = get_byte(pb);
+ }
entries = get_be32(pb);
- if(entries >= UINT_MAX / sizeof(int))
- return -1;
+
+ dprintf(c->fc, "sample_size = %d sample_count = %d\n", sc->sample_size, entries);
sc->sample_count = entries;
if (sample_size)
return 0;
- dprintf(c->fc, "sample_size = %d sample_count = %d\n", sc->sample_size, sc->sample_count);
+ if (field_size != 4 && field_size != 8 && field_size != 16 && field_size != 32) {
+ av_log(c->fc, AV_LOG_ERROR, "Invalid sample field size %d\n", field_size);
+ return -1;
+ }
+ if(entries >= UINT_MAX / sizeof(int))
+ return -1;
sc->sample_sizes = av_malloc(entries * sizeof(int));
if (!sc->sample_sizes)
+ return AVERROR(ENOMEM);
+
+ num_bytes = (entries*field_size+4)>>3;
+
+ buf = av_malloc(num_bytes+FF_INPUT_BUFFER_PADDING_SIZE);
+ if (!buf) {
+ av_freep(&sc->sample_sizes);
+ return AVERROR(ENOMEM);
+ }
+
+ if (get_buffer(pb, buf, num_bytes) < num_bytes) {
+ av_freep(&sc->sample_sizes);
+ av_free(buf);
return -1;
+ }
+
+ init_get_bits(&gb, buf, 8*num_bytes);
+
for(i=0; i<entries; i++)
- sc->sample_sizes[i] = get_be32(pb);
+ sc->sample_sizes[i] = get_bits_long(&gb, field_size);
+
+ av_free(buf);
return 0;
}
get_byte(pb); /* version */
get_be24(pb); /* flags */
entries = get_be32(pb);
+
+ dprintf(c->fc, "track[%i].stts.entries = %i\n", c->fc->nb_streams-1, entries);
+
if(entries >= UINT_MAX / sizeof(*sc->stts_data))
return -1;
-
- sc->stts_count = entries;
sc->stts_data = av_malloc(entries * sizeof(*sc->stts_data));
if (!sc->stts_data)
- return -1;
- dprintf(c->fc, "track[%i].stts.entries = %i\n", c->fc->nb_streams-1, entries);
-
- sc->time_rate=0;
+ return AVERROR(ENOMEM);
+ sc->stts_count = entries;
for(i=0; i<entries; i++) {
int sample_duration;
sc->stts_data[i].count= sample_count;
sc->stts_data[i].duration= sample_duration;
- sc->time_rate= ff_gcd(sc->time_rate, sample_duration);
+ sc->time_rate= av_gcd(sc->time_rate, sample_duration);
dprintf(c->fc, "sample_count=%d, sample_duration=%d\n",sample_count,sample_duration);
return 0;
}
+static int mov_read_cslg(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
+{
+ AVStream *st;
+ MOVStreamContext *sc;
+
+ if (c->fc->nb_streams < 1)
+ return 0;
+ st = c->fc->streams[c->fc->nb_streams-1];
+ sc = st->priv_data;
+
+ get_be32(pb); // version + flags
+
+ sc->dts_shift = get_be32(pb);
+ dprintf(c->fc, "dts shift %d\n", sc->dts_shift);
+
+ sc->time_rate= av_gcd(sc->time_rate, FFABS(sc->dts_shift));
+
+ get_be32(pb); // least dts to pts delta
+ get_be32(pb); // greatest dts to pts delta
+ get_be32(pb); // pts start
+ get_be32(pb); // pts end
+
+ return 0;
+}
+
static int mov_read_ctts(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
{
AVStream *st = c->fc->streams[c->fc->nb_streams-1];
get_byte(pb); /* version */
get_be24(pb); /* flags */
entries = get_be32(pb);
+
+ dprintf(c->fc, "track[%i].ctts.entries = %i\n", c->fc->nb_streams-1, entries);
+
if(entries >= UINT_MAX / sizeof(*sc->ctts_data))
return -1;
-
- sc->ctts_count = entries;
sc->ctts_data = av_malloc(entries * sizeof(*sc->ctts_data));
if (!sc->ctts_data)
- return -1;
- dprintf(c->fc, "track[%i].ctts.entries = %i\n", c->fc->nb_streams-1, entries);
+ return AVERROR(ENOMEM);
+ sc->ctts_count = entries;
for(i=0; i<entries; i++) {
int count =get_be32(pb);
int duration =get_be32(pb);
- if (duration < 0) {
- sc->wrong_dts = 1;
- st->codec->has_b_frames = 1;
- }
sc->ctts_data[i].count = count;
sc->ctts_data[i].duration= duration;
- sc->time_rate= ff_gcd(sc->time_rate, FFABS(duration));
+ sc->time_rate= av_gcd(sc->time_rate, FFABS(duration));
}
return 0;
}
unsigned int stss_index = 0;
unsigned int i, j;
+ /* adjust first dts according to edit list */
+ if (sc->time_offset) {
+ int rescaled = sc->time_offset < 0 ? av_rescale(sc->time_offset, sc->time_scale, mov->time_scale) : sc->time_offset;
+ assert(sc->time_offset % sc->time_rate == 0);
+ current_dts = - (rescaled / sc->time_rate);
+ if (sc->ctts_data && sc->ctts_data[0].duration / sc->stts_data[0].duration > 16) {
+ /* more than 16 frames delay, dts are likely wrong
+ this happens with files created by iMovie */
+ sc->wrong_dts = 1;
+ st->codec->has_b_frames = 1;
+ }
+ }
+
/* only use old uncompressed audio chunk demuxing when stts specifies it */
if (!(st->codec->codec_type == CODEC_TYPE_AUDIO &&
sc->stts_count == 1 && sc->stts_data[0].duration == 1)) {
unsigned int distance = 0;
int key_off = sc->keyframes && sc->keyframes[0] == 1;
+ sc->dts_shift /= sc->time_rate;
+ current_dts -= sc->dts_shift;
+
st->nb_frames = sc->sample_count;
for (i = 0; i < sc->chunk_count; i++) {
current_offset = sc->chunk_offsets[i];
- if (stsc_index + 1 < sc->sample_to_chunk_sz &&
- i + 1 == sc->sample_to_chunk[stsc_index + 1].first)
+ if (stsc_index + 1 < sc->stsc_count &&
+ i + 1 == sc->stsc_data[stsc_index + 1].first)
stsc_index++;
- for (j = 0; j < sc->sample_to_chunk[stsc_index].count; j++) {
+ for (j = 0; j < sc->stsc_data[stsc_index].count; j++) {
if (current_sample >= sc->sample_count) {
av_log(mov->fc, AV_LOG_ERROR, "wrong sample count\n");
- goto out;
+ return;
}
keyframe = !sc->keyframe_count || current_sample+key_off == sc->keyframes[stss_index];
if (keyframe) {
}
sample_size = sc->sample_size > 0 ? sc->sample_size : sc->sample_sizes[current_sample];
if(sc->pseudo_stream_id == -1 ||
- sc->sample_to_chunk[stsc_index].id - 1 == sc->pseudo_stream_id) {
+ sc->stsc_data[stsc_index].id - 1 == sc->pseudo_stream_id) {
av_add_index_entry(st, current_offset, current_dts, sample_size, distance,
keyframe ? AVINDEX_KEYFRAME : 0);
dprintf(mov->fc, "AVIndex stream %d, sample %d, offset %"PRIx64", dts %"PRId64", "
"size %d, distance %d, keyframe %d\n", st->index, current_sample,
current_offset, current_dts, sample_size, distance, keyframe);
}
+
current_offset += sample_size;
assert(sc->stts_data[stts_index].duration % sc->time_rate == 0);
current_dts += sc->stts_data[stts_index].duration / sc->time_rate;
unsigned int frames = 1;
for (i = 0; i < sc->chunk_count; i++) {
current_offset = sc->chunk_offsets[i];
- if (stsc_index + 1 < sc->sample_to_chunk_sz &&
- i + 1 == sc->sample_to_chunk[stsc_index + 1].first)
+ if (stsc_index + 1 < sc->stsc_count &&
+ i + 1 == sc->stsc_data[stsc_index + 1].first)
stsc_index++;
- chunk_samples = sc->sample_to_chunk[stsc_index].count;
+ chunk_samples = sc->stsc_data[stsc_index].count;
/* get chunk size, beware of alaw/ulaw/mace */
if (sc->samples_per_frame > 0 &&
(chunk_samples * sc->bytes_per_frame % sc->samples_per_frame == 0)) {
}
}
}
- out:
- /* adjust sample count to avindex entries */
- sc->sample_count = st->nb_index_entries;
}
static int mov_read_trak(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
st->priv_data = sc;
st->codec->codec_type = CODEC_TYPE_DATA;
- st->start_time = 0; /* XXX: check */
+ sc->ffindex = st->index;
if ((ret = mov_read_default(c, pb, atom)) < 0)
return ret;
/* sanity checks */
- if(sc->chunk_count && (!sc->stts_count || !sc->sample_to_chunk_sz ||
- (!sc->sample_size && !sc->sample_count))){
+ if (sc->chunk_count && (!sc->stts_count || !sc->stsc_count ||
+ (!sc->sample_size && !sc->sample_count))) {
av_log(c->fc, AV_LOG_ERROR, "stream %d, missing mandatory atoms, broken header\n",
st->index);
- sc->sample_count = 0; //ignore track
return 0;
}
- if(!sc->time_rate)
- sc->time_rate=1;
- if(!sc->time_scale)
- sc->time_scale= c->time_scale;
+
+ if (!sc->time_rate)
+ sc->time_rate = 1;
+ if (!sc->time_scale)
+ sc->time_scale = c->time_scale;
+
av_set_pts_info(st, 64, sc->time_rate, sc->time_scale);
if (st->codec->codec_type == CODEC_TYPE_AUDIO &&
- !st->codec->frame_size && sc->stts_count == 1)
- st->codec->frame_size = av_rescale(sc->time_rate, st->codec->sample_rate, sc->time_scale);
+ !st->codec->frame_size && sc->stts_count == 1) {
+ st->codec->frame_size = av_rescale(sc->stts_data[0].duration,
+ st->codec->sample_rate, sc->time_scale);
+ dprintf(c->fc, "frame size %d\n", st->codec->frame_size);
+ }
- if(st->duration != AV_NOPTS_VALUE){
+ if (st->duration != AV_NOPTS_VALUE) {
assert(st->duration % sc->time_rate == 0);
st->duration /= sc->time_rate;
}
- sc->ffindex = st->index;
+
mov_build_index(c, st);
if (sc->dref_id-1 < sc->drefs_count && sc->drefs[sc->dref_id-1].path) {
sc->pb = c->fc->pb;
switch (st->codec->codec_id) {
-#ifdef CONFIG_H261_DECODER
+#if CONFIG_H261_DECODER
case CODEC_ID_H261:
#endif
-#ifdef CONFIG_H263_DECODER
+#if CONFIG_H263_DECODER
case CODEC_ID_H263:
#endif
-#ifdef CONFIG_MPEG4_DECODER
+#if CONFIG_H264_DECODER
+ case CODEC_ID_H264:
+#endif
+#if CONFIG_MPEG4_DECODER
case CODEC_ID_MPEG4:
#endif
- st->codec->width= 0; /* let decoder init width/height */
+ st->codec->width = 0; /* let decoder init width/height */
st->codec->height= 0;
break;
}
/* Do not need those anymore. */
av_freep(&sc->chunk_offsets);
- av_freep(&sc->sample_to_chunk);
+ av_freep(&sc->stsc_data);
av_freep(&sc->sample_sizes);
av_freep(&sc->keyframes);
av_freep(&sc->stts_data);
static int mov_read_meta(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
{
- url_fskip(pb, 4); // version + flags
- atom.size -= 4;
- return mov_read_default(c, pb, atom);
-}
-
-static int mov_read_trkn(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
-{
- get_be32(pb); // type
- get_be32(pb); // unknown
- c->fc->track = get_be32(pb);
- dprintf(c->fc, "%.4s %d\n", (char*)&atom.type, c->fc->track);
- return 0;
-}
-
-static int mov_read_udta_string(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
-{
- char *str = NULL;
- int size;
- uint16_t str_size;
-
- if (c->itunes_metadata) {
- int data_size = get_be32(pb);
- int tag = get_le32(pb);
- if (tag == MKTAG('d','a','t','a')) {
- get_be32(pb); // type
- get_be32(pb); // unknown
- str_size = data_size - 16;
- } else return 0;
- } else {
- str_size = get_be16(pb); // string length
- get_be16(pb); // language
- }
- switch (atom.type) {
- case MKTAG(0xa9,'n','a','m'):
- str = c->fc->title; size = sizeof(c->fc->title); break;
- case MKTAG(0xa9,'A','R','T'):
- case MKTAG(0xa9,'w','r','t'):
- str = c->fc->author; size = sizeof(c->fc->author); break;
- case MKTAG(0xa9,'c','p','y'):
- str = c->fc->copyright; size = sizeof(c->fc->copyright); break;
- case MKTAG(0xa9,'c','m','t'):
- case MKTAG(0xa9,'i','n','f'):
- str = c->fc->comment; size = sizeof(c->fc->comment); break;
- case MKTAG(0xa9,'a','l','b'):
- str = c->fc->album; size = sizeof(c->fc->album); break;
+ while (atom.size > 8) {
+ uint32_t tag = get_le32(pb);
+ atom.size -= 4;
+ if (tag == MKTAG('h','d','l','r')) {
+ url_fseek(pb, -8, SEEK_CUR);
+ atom.size += 8;
+ return mov_read_default(c, pb, atom);
+ }
}
- if (!str)
- return 0;
- get_buffer(pb, str, FFMIN(size, str_size));
- dprintf(c->fc, "%.4s %s\n", (char*)&atom.type, str);
return 0;
}
int64_t disp_transform[2];
int display_matrix[3][2];
AVStream *st = c->fc->streams[c->fc->nb_streams-1];
+ MOVStreamContext *sc = st->priv_data;
int version = get_byte(pb);
get_be24(pb); /* flags */
}
st->id = (int)get_be32(pb); /* track id (NOT 0 !)*/
get_be32(pb); /* reserved */
- st->start_time = 0; /* check */
+
/* highlevel (considering edits) duration in movie timebase */
(version == 1) ? get_be64(pb) : get_be32(pb);
get_be32(pb); /* reserved */
width = get_be32(pb); // 16.16 fixed point track width
height = get_be32(pb); // 16.16 fixed point track height
+ sc->width = width >> 16;
+ sc->height = height >> 16;
//transform the display width/height according to the matrix
// skip this if the display matrix is the default identity matrix
flags = get_be24(pb);
track_id = get_be32(pb);
- if (!track_id || track_id > c->fc->nb_streams)
+ if (!track_id)
return -1;
frag->track_id = track_id;
for (i = 0; i < c->trex_count; i++)
if ((uint64_t)c->trex_count+1 >= UINT_MAX / sizeof(*c->trex_data))
return -1;
- c->trex_data = av_realloc(c->trex_data, (c->trex_count+1)*sizeof(*c->trex_data));
- if (!c->trex_data)
+ trex = av_realloc(c->trex_data, (c->trex_count+1)*sizeof(*c->trex_data));
+ if (!trex)
return AVERROR(ENOMEM);
+ c->trex_data = trex;
trex = &c->trex_data[c->trex_count++];
get_byte(pb); /* version */
get_be24(pb); /* flags */
static int mov_read_trun(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
{
MOVFragment *frag = &c->fragment;
- AVStream *st;
+ AVStream *st = NULL;
MOVStreamContext *sc;
uint64_t offset;
int64_t dts;
unsigned entries, first_sample_flags = frag->flags;
int flags, distance, i;
- if (!frag->track_id || frag->track_id > c->fc->nb_streams)
+ for (i = 0; i < c->fc->nb_streams; i++) {
+ if (c->fc->streams[i]->id == frag->track_id) {
+ st = c->fc->streams[i];
+ break;
+ }
+ }
+ if (!st) {
+ av_log(c->fc, AV_LOG_ERROR, "could not find corresponding track id %d\n", frag->track_id);
return -1;
- st = c->fc->streams[frag->track_id-1];
+ }
sc = st->priv_data;
if (sc->pseudo_stream_id+1 != frag->stsd_id)
return 0;
if (flags & 0x001) data_offset = get_be32(pb);
if (flags & 0x004) first_sample_flags = get_be32(pb);
if (flags & 0x800) {
+ MOVStts *ctts_data;
if ((uint64_t)entries+sc->ctts_count >= UINT_MAX/sizeof(*sc->ctts_data))
return -1;
- sc->ctts_data = av_realloc(sc->ctts_data,
- (entries+sc->ctts_count)*sizeof(*sc->ctts_data));
- if (!sc->ctts_data)
+ ctts_data = av_realloc(sc->ctts_data,
+ (entries+sc->ctts_count)*sizeof(*sc->ctts_data));
+ if (!ctts_data)
return AVERROR(ENOMEM);
+ sc->ctts_data = ctts_data;
}
dts = st->duration;
offset = frag->base_data_offset + data_offset;
offset += sample_size;
}
frag->moof_offset = offset;
- sc->sample_count = st->nb_index_entries;
st->duration = dts;
return 0;
}
static int mov_read_cmov(MOVContext *c, ByteIOContext *pb, MOVAtom atom)
{
-#ifdef CONFIG_ZLIB
+#if CONFIG_ZLIB
ByteIOContext ctx;
uint8_t *cmov_data;
uint8_t *moov_data; /* uncompressed data */
if (get_le32(pb) != MKTAG('d','c','o','m'))
return -1;
if (get_le32(pb) != MKTAG('z','l','i','b')) {
- av_log(NULL, AV_LOG_ERROR, "unknown compression for cmov atom !");
+ av_log(c->fc, AV_LOG_ERROR, "unknown compression for cmov atom !");
return -1;
}
get_be32(pb); /* cmvd atom */
cmov_data = av_malloc(cmov_len);
if (!cmov_data)
- return -1;
+ return AVERROR(ENOMEM);
moov_data = av_malloc(moov_len);
if (!moov_data) {
av_free(cmov_data);
- return -1;
+ return AVERROR(ENOMEM);
}
get_buffer(pb, cmov_data, cmov_len);
if(uncompress (moov_data, (uLongf *) &moov_len, (const Bytef *)cmov_data, cmov_len) != Z_OK)
get_byte(pb); /* version */
get_be24(pb); /* flags */
- edit_count= sc->edit_count = get_be32(pb); /* entries */
+ edit_count = get_be32(pb); /* entries */
for(i=0; i<edit_count; i++){
int time;
- get_be32(pb); /* Track duration */
+ int duration = get_be32(pb); /* Track duration */
time = get_be32(pb); /* Media time */
get_be32(pb); /* Media rate */
- if (time != 0)
- av_log(c->fc, AV_LOG_WARNING, "edit list not starting at 0, "
- "a/v desync might occur, patch welcome\n");
+ if (i == 0 && time >= -1) {
+ sc->time_offset = time != -1 ? time : -duration;
+ sc->time_rate = av_gcd(sc->time_rate, FFABS(sc->time_offset));
+ }
}
- dprintf(c->fc, "track[%i].edit_count = %i\n", c->fc->nb_streams-1, sc->edit_count);
+
+ if(edit_count > 1)
+ av_log(c->fc, AV_LOG_WARNING, "multiple edit list entries, "
+ "a/v desync might occur, patch welcome\n");
+
+ dprintf(c->fc, "track[%i].edit_count = %i\n", c->fc->nb_streams-1, edit_count);
return 0;
}
static const MOVParseTableEntry mov_default_parse_table[] = {
{ MKTAG('a','v','s','s'), mov_read_extradata },
{ MKTAG('c','o','6','4'), mov_read_stco },
+{ MKTAG('c','s','l','g'), mov_read_cslg },
{ MKTAG('c','t','t','s'), mov_read_ctts }, /* composition time to sample */
{ MKTAG('d','i','n','f'), mov_read_default },
{ MKTAG('d','r','e','f'), mov_read_dref },
{ MKTAG('s','t','s','s'), mov_read_stss }, /* sync sample */
{ MKTAG('s','t','s','z'), mov_read_stsz }, /* sample size */
{ MKTAG('s','t','t','s'), mov_read_stts },
+{ MKTAG('s','t','z','2'), mov_read_stsz }, /* compact sample size */
{ MKTAG('t','k','h','d'), mov_read_tkhd }, /* track header */
{ MKTAG('t','f','h','d'), mov_read_tfhd }, /* track fragment header */
{ MKTAG('t','r','a','k'), mov_read_trak },
{ MKTAG('t','r','a','f'), mov_read_default },
{ MKTAG('t','r','e','x'), mov_read_trex },
-{ MKTAG('t','r','k','n'), mov_read_trkn },
{ MKTAG('t','r','u','n'), mov_read_trun },
{ MKTAG('u','d','t','a'), mov_read_default },
{ MKTAG('w','a','v','e'), mov_read_wave },
{ MKTAG('e','s','d','s'), mov_read_esds },
{ MKTAG('w','i','d','e'), mov_read_wide }, /* place holder */
{ MKTAG('c','m','o','v'), mov_read_cmov },
-{ MKTAG(0xa9,'n','a','m'), mov_read_udta_string },
-{ MKTAG(0xa9,'w','r','t'), mov_read_udta_string },
-{ MKTAG(0xa9,'c','p','y'), mov_read_udta_string },
-{ MKTAG(0xa9,'i','n','f'), mov_read_udta_string },
-{ MKTAG(0xa9,'i','n','f'), mov_read_udta_string },
-{ MKTAG(0xa9,'A','R','T'), mov_read_udta_string },
-{ MKTAG(0xa9,'a','l','b'), mov_read_udta_string },
-{ MKTAG(0xa9,'c','m','t'), mov_read_udta_string },
{ 0, NULL }
};
MOVStreamContext *sc = 0;
AVIndexEntry *sample = 0;
int64_t best_dts = INT64_MAX;
- int i;
+ int i, ret;
retry:
for (i = 0; i < s->nb_streams; i++) {
AVStream *st = s->streams[i];
MOVStreamContext *msc = st->priv_data;
- if (st->discard != AVDISCARD_ALL && msc->pb && msc->current_sample < msc->sample_count) {
+ if (st->discard != AVDISCARD_ALL && msc->pb && msc->current_sample < st->nb_index_entries) {
AVIndexEntry *current_sample = &st->index_entries[msc->current_sample];
int64_t dts = av_rescale(current_sample->timestamp * (int64_t)msc->time_rate,
AV_TIME_BASE, msc->time_scale);
sc->ffindex, sample->pos);
return -1;
}
- av_get_packet(sc->pb, pkt, sample->size);
-#ifdef CONFIG_DV_DEMUXER
+ ret = av_get_packet(sc->pb, pkt, sample->size);
+ if (ret < 0)
+ return ret;
+#if CONFIG_DV_DEMUXER
if (mov->dv_demux && sc->dv_audio_container) {
dv_produce_packet(mov->dv_demux, pkt, pkt->data, pkt->size);
av_free(pkt->data);
pkt->stream_index = sc->ffindex;
pkt->dts = sample->timestamp;
if (sc->ctts_data) {
- assert(sc->ctts_data[sc->sample_to_ctime_index].duration % sc->time_rate == 0);
- pkt->pts = pkt->dts + sc->ctts_data[sc->sample_to_ctime_index].duration / sc->time_rate;
+ assert(sc->ctts_data[sc->ctts_index].duration % sc->time_rate == 0);
+ pkt->pts = pkt->dts + sc->dts_shift + sc->ctts_data[sc->ctts_index].duration / sc->time_rate;
/* update ctts context */
- sc->sample_to_ctime_sample++;
- if (sc->sample_to_ctime_index < sc->ctts_count &&
- sc->ctts_data[sc->sample_to_ctime_index].count == sc->sample_to_ctime_sample) {
- sc->sample_to_ctime_index++;
- sc->sample_to_ctime_sample = 0;
+ sc->ctts_sample++;
+ if (sc->ctts_index < sc->ctts_count &&
+ sc->ctts_data[sc->ctts_index].count == sc->ctts_sample) {
+ sc->ctts_index++;
+ sc->ctts_sample = 0;
}
if (sc->wrong_dts)
pkt->dts = AV_NOPTS_VALUE;
} else {
AVStream *st = s->streams[sc->ffindex];
- int64_t next_dts = (sc->current_sample < sc->sample_count) ?
+ int64_t next_dts = (sc->current_sample < st->nb_index_entries) ?
st->index_entries[sc->current_sample].timestamp : st->duration;
pkt->duration = next_dts - pkt->dts;
pkt->pts = pkt->dts;
for (i = 0; i < sc->ctts_count; i++) {
int next = time_sample + sc->ctts_data[i].count;
if (next > sc->current_sample) {
- sc->sample_to_ctime_index = i;
- sc->sample_to_ctime_sample = sc->current_sample - time_sample;
+ sc->ctts_index = i;
+ sc->ctts_sample = sc->current_sample - time_sample;
break;
}
time_sample = next;
if (stream_index >= s->nb_streams)
return -1;
+ if (sample_time < 0)
+ sample_time = 0;
st = s->streams[stream_index];
sample = mov_seek_stream(st, sample_time, flags);
static int mov_read_close(AVFormatContext *s)
{
- int i, j;
MOVContext *mov = s->priv_data;
- for(i=0; i<s->nb_streams; i++) {
- MOVStreamContext *sc = s->streams[i]->priv_data;
+ int i, j;
+
+ for (i = 0; i < s->nb_streams; i++) {
+ AVStream *st = s->streams[i];
+ MOVStreamContext *sc = st->priv_data;
+
av_freep(&sc->ctts_data);
- for (j=0; j<sc->drefs_count; j++)
+ for (j = 0; j < sc->drefs_count; j++)
av_freep(&sc->drefs[j].path);
av_freep(&sc->drefs);
if (sc->pb && sc->pb != s->pb)
url_fclose(sc->pb);
+
+ av_freep(&st->codec->palctrl);
}
- if(mov->dv_demux){
- for(i=0; i<mov->dv_fctx->nb_streams; i++){
+
+ if (mov->dv_demux) {
+ for(i = 0; i < mov->dv_fctx->nb_streams; i++) {
av_freep(&mov->dv_fctx->streams[i]->codec);
av_freep(&mov->dv_fctx->streams[i]);
}
av_freep(&mov->dv_fctx);
av_freep(&mov->dv_demux);
}
+
av_freep(&mov->trex_data);
+
return 0;
}
projects / frescor / ffmpeg.git / blobdiff