/*
- * Matroska file demuxer (no muxer yet)
- * Copyright (c) 2003-2004 The ffmpeg Project
+ * Matroska file demuxer
+ * Copyright (c) 2003-2008 The FFmpeg Project
*
* This file is part of FFmpeg.
*
*/
/**
- * @file matroskadec.c
+ * @file libavformat/matroskadec.c
* Matroska file demuxer
* by Ronald Bultje <rbultje@ronald.bitfreak.net>
* with a little help from Moritz Bunkus <moritz@bunkus.org>
- * Specs available on the matroska project page:
- * http://www.matroska.org/.
+ * totally reworked by Aurelien Jacobs <aurel@gnuage.org>
+ * Specs available on the Matroska project page: http://www.matroska.org/.
*/
+#include <stdio.h>
#include "avformat.h"
-/* For codec_get_id(). */
+/* For ff_codec_get_id(). */
#include "riff.h"
#include "isom.h"
#include "matroska.h"
#include "libavcodec/mpeg4audio.h"
#include "libavutil/intfloat_readwrite.h"
+#include "libavutil/intreadwrite.h"
#include "libavutil/avstring.h"
#include "libavutil/lzo.h"
-#ifdef CONFIG_ZLIB
+#if CONFIG_ZLIB
#include <zlib.h>
#endif
-#ifdef CONFIG_BZLIB
+#if CONFIG_BZLIB
#include <bzlib.h>
#endif
typedef struct {
uint64_t num;
+ uint64_t uid;
uint64_t type;
+ char *name;
char *codec_id;
EbmlBin codec_priv;
char *language;
EbmlList encodings;
AVStream *stream;
+ int64_t end_timecode;
+ int ms_compat;
} MatroskaTrack;
typedef struct {
+ uint64_t uid;
char *filename;
char *mime;
EbmlBin bin;
+
+ AVStream *stream;
} MatroskaAttachement;
typedef struct {
uint64_t end;
uint64_t uid;
char *title;
+
+ AVChapter *chapter;
} MatroskaChapter;
typedef struct {
EbmlList pos;
} MatroskaIndex;
+typedef struct {
+ char *name;
+ char *string;
+ char *lang;
+ uint64_t def;
+ EbmlList sub;
+} MatroskaTag;
+
+typedef struct {
+ char *type;
+ uint64_t typevalue;
+ uint64_t trackuid;
+ uint64_t chapteruid;
+ uint64_t attachuid;
+} MatroskaTagTarget;
+
+typedef struct {
+ MatroskaTagTarget target;
+ EbmlList tag;
+} MatroskaTags;
+
typedef struct {
uint64_t id;
uint64_t pos;
typedef struct {
AVFormatContext *ctx;
- /* ebml stuff */
+ /* EBML stuff */
int num_levels;
MatroskaLevel levels[EBML_MAX_DEPTH];
int level_up;
EbmlList attachments;
EbmlList chapters;
EbmlList index;
+ EbmlList tags;
EbmlList seekhead;
- /* num_streams is the number of streams that av_new_stream() was called
- * for ( = that are available to the calling program). */
- int num_streams;
-
- /* cache for ID peeking */
- uint32_t peek_id;
-
/* byte position of the segment inside the stream */
- offset_t segment_start;
+ int64_t segment_start;
- /* The packet queue. */
+ /* the packet queue */
AVPacket **packets;
int num_packets;
+ AVPacket *prev_pkt;
int done;
int has_cluster_id;
/* What to skip before effectively reading a packet. */
int skip_to_keyframe;
- AVStream *skip_to_stream;
+ uint64_t skip_to_timecode;
} MatroskaDemuxContext;
typedef struct {
uint64_t duration;
int64_t reference;
+ uint64_t non_simple;
EbmlBin bin;
} MatroskaBlock;
EbmlList blocks;
} MatroskaCluster;
-#define ARRAY_SIZE(x) (sizeof(x)/sizeof(*x))
-
static EbmlSyntax ebml_header[] = {
{ EBML_ID_EBMLREADVERSION, EBML_UINT, 0, offsetof(Ebml,version), {.u=EBML_VERSION} },
{ EBML_ID_EBMLMAXSIZELENGTH, EBML_UINT, 0, offsetof(Ebml,max_size), {.u=8} },
{ EBML_ID_DOCTYPEREADVERSION, EBML_UINT, 0, offsetof(Ebml,doctype_version), {.u=1} },
{ EBML_ID_EBMLVERSION, EBML_NONE },
{ EBML_ID_DOCTYPEVERSION, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_MUXINGAPP, EBML_NONE },
{ MATROSKA_ID_DATEUTC, EBML_NONE },
{ MATROSKA_ID_SEGMENTUID, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_VIDEOPIXELWIDTH, EBML_UINT, 0, offsetof(MatroskaTrackVideo,pixel_width) },
{ MATROSKA_ID_VIDEOPIXELHEIGHT, EBML_UINT, 0, offsetof(MatroskaTrackVideo,pixel_height) },
{ MATROSKA_ID_VIDEOCOLORSPACE, EBML_UINT, 0, offsetof(MatroskaTrackVideo,fourcc) },
+ { MATROSKA_ID_VIDEOPIXELCROPB, EBML_NONE },
+ { MATROSKA_ID_VIDEOPIXELCROPT, EBML_NONE },
+ { MATROSKA_ID_VIDEOPIXELCROPL, EBML_NONE },
+ { MATROSKA_ID_VIDEOPIXELCROPR, EBML_NONE },
+ { MATROSKA_ID_VIDEODISPLAYUNIT, EBML_NONE },
{ MATROSKA_ID_VIDEOFLAGINTERLACED,EBML_NONE },
{ MATROSKA_ID_VIDEOSTEREOMODE, EBML_NONE },
{ MATROSKA_ID_VIDEOASPECTRATIO, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_AUDIOOUTSAMPLINGFREQ,EBML_FLOAT,0,offsetof(MatroskaTrackAudio,out_samplerate) },
{ MATROSKA_ID_AUDIOBITDEPTH, EBML_UINT, 0, offsetof(MatroskaTrackAudio,bitdepth) },
{ MATROSKA_ID_AUDIOCHANNELS, EBML_UINT, 0, offsetof(MatroskaTrackAudio,channels), {.u=1} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_track_encoding_compression[] = {
{ MATROSKA_ID_ENCODINGCOMPALGO, EBML_UINT, 0, offsetof(MatroskaTrackCompression,algo), {.u=0} },
{ MATROSKA_ID_ENCODINGCOMPSETTINGS,EBML_BIN, 0, offsetof(MatroskaTrackCompression,settings) },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_ENCODINGSCOPE, EBML_UINT, 0, offsetof(MatroskaTrackEncoding,scope), {.u=1} },
{ MATROSKA_ID_ENCODINGTYPE, EBML_UINT, 0, offsetof(MatroskaTrackEncoding,type), {.u=0} },
{ MATROSKA_ID_ENCODINGCOMPRESSION,EBML_NEST, 0, offsetof(MatroskaTrackEncoding,compression), {.n=matroska_track_encoding_compression} },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_ENCODINGORDER, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_track_encodings[] = {
{ MATROSKA_ID_TRACKCONTENTENCODING, EBML_NEST, sizeof(MatroskaTrackEncoding), offsetof(MatroskaTrack,encodings), {.n=matroska_track_encoding} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_track[] = {
{ MATROSKA_ID_TRACKNUMBER, EBML_UINT, 0, offsetof(MatroskaTrack,num) },
+ { MATROSKA_ID_TRACKNAME, EBML_UTF8, 0, offsetof(MatroskaTrack,name) },
+ { MATROSKA_ID_TRACKUID, EBML_UINT, 0, offsetof(MatroskaTrack,uid) },
{ MATROSKA_ID_TRACKTYPE, EBML_UINT, 0, offsetof(MatroskaTrack,type) },
{ MATROSKA_ID_CODECID, EBML_STR, 0, offsetof(MatroskaTrack,codec_id) },
{ MATROSKA_ID_CODECPRIVATE, EBML_BIN, 0, offsetof(MatroskaTrack,codec_priv) },
{ MATROSKA_ID_TRACKVIDEO, EBML_NEST, 0, offsetof(MatroskaTrack,video), {.n=matroska_track_video} },
{ MATROSKA_ID_TRACKAUDIO, EBML_NEST, 0, offsetof(MatroskaTrack,audio), {.n=matroska_track_audio} },
{ MATROSKA_ID_TRACKCONTENTENCODINGS,EBML_NEST, 0, 0, {.n=matroska_track_encodings} },
- { MATROSKA_ID_TRACKUID, EBML_NONE },
- { MATROSKA_ID_TRACKNAME, EBML_NONE },
{ MATROSKA_ID_TRACKFLAGENABLED, EBML_NONE },
{ MATROSKA_ID_TRACKFLAGFORCED, EBML_NONE },
{ MATROSKA_ID_TRACKFLAGLACING, EBML_NONE },
{ MATROSKA_ID_CODECDOWNLOADURL, EBML_NONE },
{ MATROSKA_ID_TRACKMINCACHE, EBML_NONE },
{ MATROSKA_ID_TRACKMAXCACHE, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_TRACKMAXBLKADDID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_tracks[] = {
{ MATROSKA_ID_TRACKENTRY, EBML_NEST, sizeof(MatroskaTrack), offsetof(MatroskaDemuxContext,tracks), {.n=matroska_track} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_attachment[] = {
+ { MATROSKA_ID_FILEUID, EBML_UINT, 0, offsetof(MatroskaAttachement,uid) },
{ MATROSKA_ID_FILENAME, EBML_UTF8, 0, offsetof(MatroskaAttachement,filename) },
{ MATROSKA_ID_FILEMIMETYPE, EBML_STR, 0, offsetof(MatroskaAttachement,mime) },
{ MATROSKA_ID_FILEDATA, EBML_BIN, 0, offsetof(MatroskaAttachement,bin) },
- { MATROSKA_ID_FILEUID, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_FILEDESC, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_attachments[] = {
{ MATROSKA_ID_ATTACHEDFILE, EBML_NEST, sizeof(MatroskaAttachement), offsetof(MatroskaDemuxContext,attachments), {.n=matroska_attachment} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_chapter_display[] = {
{ MATROSKA_ID_CHAPSTRING, EBML_UTF8, 0, offsetof(MatroskaChapter,title) },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_CHAPLANG, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_CHAPTERUID, EBML_UINT, 0, offsetof(MatroskaChapter,uid) },
{ MATROSKA_ID_CHAPTERDISPLAY, EBML_NEST, 0, 0, {.n=matroska_chapter_display} },
{ MATROSKA_ID_CHAPTERFLAGHIDDEN, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_CHAPTERFLAGENABLED, EBML_NONE },
+ { MATROSKA_ID_CHAPTERPHYSEQUIV, EBML_NONE },
+ { MATROSKA_ID_CHAPTERATOM, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_EDITIONUID, EBML_NONE },
{ MATROSKA_ID_EDITIONFLAGHIDDEN, EBML_NONE },
{ MATROSKA_ID_EDITIONFLAGDEFAULT, EBML_NONE },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_EDITIONFLAGORDERED, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_chapters[] = {
{ MATROSKA_ID_EDITIONENTRY, EBML_NEST, 0, 0, {.n=matroska_chapter} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_index_pos[] = {
{ MATROSKA_ID_CUETRACK, EBML_UINT, 0, offsetof(MatroskaIndexPos,track) },
{ MATROSKA_ID_CUECLUSTERPOSITION, EBML_UINT, 0, offsetof(MatroskaIndexPos,pos) },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_CUEBLOCKNUMBER, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_index_entry[] = {
{ MATROSKA_ID_CUETIME, EBML_UINT, 0, offsetof(MatroskaIndex,time) },
{ MATROSKA_ID_CUETRACKPOSITION, EBML_NEST, sizeof(MatroskaIndexPos), offsetof(MatroskaIndex,pos), {.n=matroska_index_pos} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_index[] = {
{ MATROSKA_ID_POINTENTRY, EBML_NEST, sizeof(MatroskaIndex), offsetof(MatroskaDemuxContext,index), {.n=matroska_index_entry} },
- { EBML_ID_VOID, EBML_NONE },
+ { 0 }
+};
+
+static EbmlSyntax matroska_simpletag[] = {
+ { MATROSKA_ID_TAGNAME, EBML_UTF8, 0, offsetof(MatroskaTag,name) },
+ { MATROSKA_ID_TAGSTRING, EBML_UTF8, 0, offsetof(MatroskaTag,string) },
+ { MATROSKA_ID_TAGLANG, EBML_STR, 0, offsetof(MatroskaTag,lang), {.s="und"} },
+ { MATROSKA_ID_TAGDEFAULT, EBML_UINT, 0, offsetof(MatroskaTag,def) },
+ { MATROSKA_ID_SIMPLETAG, EBML_NEST, sizeof(MatroskaTag), offsetof(MatroskaTag,sub), {.n=matroska_simpletag} },
+ { 0 }
+};
+
+static EbmlSyntax matroska_tagtargets[] = {
+ { MATROSKA_ID_TAGTARGETS_TYPE, EBML_STR, 0, offsetof(MatroskaTagTarget,type) },
+ { MATROSKA_ID_TAGTARGETS_TYPEVALUE, EBML_UINT, 0, offsetof(MatroskaTagTarget,typevalue), {.u=50} },
+ { MATROSKA_ID_TAGTARGETS_TRACKUID, EBML_UINT, 0, offsetof(MatroskaTagTarget,trackuid) },
+ { MATROSKA_ID_TAGTARGETS_CHAPTERUID,EBML_UINT, 0, offsetof(MatroskaTagTarget,chapteruid) },
+ { MATROSKA_ID_TAGTARGETS_ATTACHUID, EBML_UINT, 0, offsetof(MatroskaTagTarget,attachuid) },
+ { 0 }
+};
+
+static EbmlSyntax matroska_tag[] = {
+ { MATROSKA_ID_SIMPLETAG, EBML_NEST, sizeof(MatroskaTag), offsetof(MatroskaTags,tag), {.n=matroska_simpletag} },
+ { MATROSKA_ID_TAGTARGETS, EBML_NEST, 0, offsetof(MatroskaTags,target), {.n=matroska_tagtargets} },
{ 0 }
};
static EbmlSyntax matroska_tags[] = {
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_TAG, EBML_NEST, sizeof(MatroskaTags), offsetof(MatroskaDemuxContext,tags), {.n=matroska_tag} },
{ 0 }
};
static EbmlSyntax matroska_seekhead_entry[] = {
{ MATROSKA_ID_SEEKID, EBML_UINT, 0, offsetof(MatroskaSeekhead,id) },
{ MATROSKA_ID_SEEKPOSITION, EBML_UINT, 0, offsetof(MatroskaSeekhead,pos), {.u=-1} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_seekhead[] = {
{ MATROSKA_ID_SEEKENTRY, EBML_NEST, sizeof(MatroskaSeekhead), offsetof(MatroskaDemuxContext,seekhead), {.n=matroska_seekhead_entry} },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_TAGS, EBML_NEST, 0, 0, {.n=matroska_tags } },
{ MATROSKA_ID_SEEKHEAD, EBML_NEST, 0, 0, {.n=matroska_seekhead } },
{ MATROSKA_ID_CLUSTER, EBML_STOP, 0, offsetof(MatroskaDemuxContext,has_cluster_id) },
- { EBML_ID_VOID, EBML_NONE },
{ 0 }
};
{ MATROSKA_ID_SIMPLEBLOCK, EBML_BIN, 0, offsetof(MatroskaBlock,bin) },
{ MATROSKA_ID_BLOCKDURATION, EBML_UINT, 0, offsetof(MatroskaBlock,duration), {.u=AV_NOPTS_VALUE} },
{ MATROSKA_ID_BLOCKREFERENCE, EBML_UINT, 0, offsetof(MatroskaBlock,reference) },
- { EBML_ID_VOID, EBML_NONE },
+ { 1, EBML_UINT, 0, offsetof(MatroskaBlock,non_simple), {.u=1} },
{ 0 }
};
{ MATROSKA_ID_CLUSTERTIMECODE,EBML_UINT,0, offsetof(MatroskaCluster,timecode) },
{ MATROSKA_ID_BLOCKGROUP, EBML_NEST, sizeof(MatroskaBlock), offsetof(MatroskaCluster,blocks), {.n=matroska_blockgroup} },
{ MATROSKA_ID_SIMPLEBLOCK, EBML_PASS, sizeof(MatroskaBlock), offsetof(MatroskaCluster,blocks), {.n=matroska_blockgroup} },
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_CLUSTERPOSITION,EBML_NONE },
+ { MATROSKA_ID_CLUSTERPREVSIZE,EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_clusters[] = {
{ MATROSKA_ID_CLUSTER, EBML_NEST, 0, 0, {.n=matroska_cluster} },
+ { MATROSKA_ID_INFO, EBML_NONE },
+ { MATROSKA_ID_CUES, EBML_NONE },
+ { MATROSKA_ID_TAGS, EBML_NONE },
+ { MATROSKA_ID_SEEKHEAD, EBML_NONE },
{ 0 }
};
/*
- * Return: the amount of levels in the hierarchy that the
- * current element lies higher than the previous one.
- * The opposite isn't done - that's auto-done using master
- * element reading.
+ * Return: Whether we reached the end of a level in the hierarchy or not.
*/
-static int
-ebml_read_element_level_up (MatroskaDemuxContext *matroska)
+static int ebml_level_end(MatroskaDemuxContext *matroska)
{
ByteIOContext *pb = matroska->ctx->pb;
- offset_t pos = url_ftell(pb);
- int num = 0;
+ int64_t pos = url_ftell(pb);
- while (matroska->num_levels > 0) {
+ if (matroska->num_levels > 0) {
MatroskaLevel *level = &matroska->levels[matroska->num_levels - 1];
-
- if (pos >= level->start + level->length) {
+ if (pos - level->start >= level->length) {
matroska->num_levels--;
- num++;
- } else {
- break;
+ return 1;
}
}
-
- return num;
+ return 0;
}
/*
* number of 0-bits followed by a one. The position of the first
* "one" bit inside the first byte indicates the length of this
* number.
- * Returns: num. of bytes read. < 0 on error.
+ * Returns: number of bytes read, < 0 on error
*/
-static int
-ebml_read_num (MatroskaDemuxContext *matroska,
- int max_size,
- uint64_t *number)
+static int ebml_read_num(MatroskaDemuxContext *matroska, ByteIOContext *pb,
+ int max_size, uint64_t *number)
{
- ByteIOContext *pb = matroska->ctx->pb;
int len_mask = 0x80, read = 1, n = 1;
int64_t total = 0;
- /* the first byte tells us the length in bytes - get_byte() can normally
+ /* The first byte tells us the length in bytes - get_byte() can normally
* return 0, but since that's not a valid first ebmlID byte, we can
* use it safely here to catch EOS. */
if (!(total = get_byte(pb))) {
/* we might encounter EOS here */
if (!url_feof(pb)) {
- offset_t pos = url_ftell(pb);
+ int64_t pos = url_ftell(pb);
av_log(matroska->ctx, AV_LOG_ERROR,
"Read error at pos. %"PRIu64" (0x%"PRIx64")\n",
pos, pos);
len_mask >>= 1;
}
if (read > max_size) {
- offset_t pos = url_ftell(pb) - 1;
+ int64_t pos = url_ftell(pb) - 1;
av_log(matroska->ctx, AV_LOG_ERROR,
"Invalid EBML number size tag 0x%02x at pos %"PRIu64" (0x%"PRIx64")\n",
(uint8_t) total, pos, pos);
return read;
}
-/*
- * Read: the element content data ID.
- * Return: the number of bytes read or < 0 on error.
- */
-static int
-ebml_read_element_id (MatroskaDemuxContext *matroska,
- uint32_t *id)
-{
- int read;
- uint64_t total;
-
- /* if we re-call this, use our cached ID */
- if (matroska->peek_id != 0) {
- *id = matroska->peek_id;
- return 0;
- }
-
- /* read out the "EBML number", include tag in ID */
- if ((read = ebml_read_num(matroska, 4, &total)) < 0)
- return read;
- *id = matroska->peek_id = total | (1 << (read * 7));
-
- return read;
-}
-
-/*
- * Read: element content length.
- * Return: the number of bytes read or < 0 on error.
- */
-static int
-ebml_read_element_length (MatroskaDemuxContext *matroska,
- uint64_t *length)
-{
- /* clear cache since we're now beyond that data point */
- matroska->peek_id = 0;
-
- /* read out the "EBML number", include tag in ID */
- return ebml_read_num(matroska, 8, length);
-}
-
-/*
- * Seek to a given offset.
- * 0 is success, -1 is failure.
- */
-static int
-ebml_read_seek (MatroskaDemuxContext *matroska,
- offset_t offset)
-{
- ByteIOContext *pb = matroska->ctx->pb;
-
- /* clear ID cache, if any */
- matroska->peek_id = 0;
-
- return (url_fseek(pb, offset, SEEK_SET) == offset) ? 0 : -1;
-}
-
/*
* Read the next element as an unsigned int.
* 0 is success, < 0 is failure.
*/
-static int
-ebml_read_uint (ByteIOContext *pb,
- int size,
- uint64_t *num)
+static int ebml_read_uint(ByteIOContext *pb, int size, uint64_t *num)
{
int n = 0;
if (size < 1 || size > 8)
return AVERROR_INVALIDDATA;
- /* big-endian ordening; build up number */
+ /* big-endian ordering; build up number */
*num = 0;
while (n++ < size)
*num = (*num << 8) | get_byte(pb);
* Read the next element as a float.
* 0 is success, < 0 is failure.
*/
-static int
-ebml_read_float (ByteIOContext *pb,
- int size,
- double *num)
+static int ebml_read_float(ByteIOContext *pb, int size, double *num)
{
if (size == 4) {
*num= av_int2flt(get_be32(pb));
* Read the next element as an ASCII string.
* 0 is success, < 0 is failure.
*/
-static int
-ebml_read_ascii (ByteIOContext *pb,
- int size,
- char **str)
+static int ebml_read_ascii(ByteIOContext *pb, int size, char **str)
{
av_free(*str);
- /* ebml strings are usually not 0-terminated, so we allocate one
+ /* EBML strings are usually not 0-terminated, so we allocate one
* byte more, read the string and NULL-terminate it ourselves. */
if (!(*str = av_malloc(size + 1)))
return AVERROR(ENOMEM);
return 0;
}
+/*
+ * Read the next element as binary data.
+ * 0 is success, < 0 is failure.
+ */
+static int ebml_read_binary(ByteIOContext *pb, int length, EbmlBin *bin)
+{
+ av_free(bin->data);
+ if (!(bin->data = av_malloc(length)))
+ return AVERROR(ENOMEM);
+
+ bin->size = length;
+ bin->pos = url_ftell(pb);
+ if (get_buffer(pb, bin->data, length) != length)
+ return AVERROR(EIO);
+
+ return 0;
+}
+
/*
* Read the next element, but only the header. The contents
* are supposed to be sub-elements which can be read separately.
* 0 is success, < 0 is failure.
*/
-static int
-ebml_read_master (MatroskaDemuxContext *matroska,
- int length)
+static int ebml_read_master(MatroskaDemuxContext *matroska, int length)
{
ByteIOContext *pb = matroska->ctx->pb;
MatroskaLevel *level;
return 0;
}
-/*
- * Read the next element as binary data.
- * 0 is success, < 0 is failure.
- */
-static int
-ebml_read_binary (ByteIOContext *pb,
- int length,
- EbmlBin *bin)
-{
- av_free(bin->data);
- if (!(bin->data = av_malloc(length)))
- return AVERROR(ENOMEM);
-
- bin->size = length;
- bin->pos = url_ftell(pb);
- if (get_buffer(pb, bin->data, length) != length)
- return AVERROR(EIO);
-
- return 0;
-}
-
/*
* Read signed/unsigned "EBML" numbers.
- * Return: number of bytes processed, < 0 on error.
- * XXX: use ebml_read_num().
+ * Return: number of bytes processed, < 0 on error
*/
-static int
-matroska_ebmlnum_uint (uint8_t *data,
- uint32_t size,
- uint64_t *num)
+static int matroska_ebmlnum_uint(MatroskaDemuxContext *matroska,
+ uint8_t *data, uint32_t size, uint64_t *num)
{
- int len_mask = 0x80, read = 1, n = 1, num_ffs = 0;
- uint64_t total;
-
- if (size <= 0)
- return AVERROR_INVALIDDATA;
-
- total = data[0];
- while (read <= 8 && !(total & len_mask)) {
- read++;
- len_mask >>= 1;
- }
- if (read > 8)
- return AVERROR_INVALIDDATA;
-
- if ((total &= (len_mask - 1)) == len_mask - 1)
- num_ffs++;
- if (size < read)
- return AVERROR_INVALIDDATA;
- while (n < read) {
- if (data[n] == 0xff)
- num_ffs++;
- total = (total << 8) | data[n];
- n++;
- }
-
- if (read == num_ffs)
- *num = (uint64_t)-1;
- else
- *num = total;
-
- return read;
+ ByteIOContext pb;
+ init_put_byte(&pb, data, size, 0, NULL, NULL, NULL, NULL);
+ return ebml_read_num(matroska, &pb, 8, num);
}
/*
* Same as above, but signed.
*/
-static int
-matroska_ebmlnum_sint (uint8_t *data,
- uint32_t size,
- int64_t *num)
+static int matroska_ebmlnum_sint(MatroskaDemuxContext *matroska,
+ uint8_t *data, uint32_t size, int64_t *num)
{
uint64_t unum;
int res;
/* read as unsigned number first */
- if ((res = matroska_ebmlnum_uint(data, size, &unum)) < 0)
+ if ((res = matroska_ebmlnum_uint(matroska, data, size, &unum)) < 0)
return res;
/* make signed (weird way) */
- if (unum == (uint64_t)-1)
- *num = INT64_MAX;
- else
- *num = unum - ((1LL << ((7 * res) - 1)) - 1);
+ *num = unum - ((1LL << (7*res - 1)) - 1);
return res;
}
+static int ebml_parse_elem(MatroskaDemuxContext *matroska,
+ EbmlSyntax *syntax, void *data);
-static MatroskaTrack *
-matroska_find_track_by_num (MatroskaDemuxContext *matroska,
- int num)
+static int ebml_parse_id(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
+ uint32_t id, void *data)
{
- MatroskaTrack *tracks = matroska->tracks.elem;
int i;
-
- for (i=0; i < matroska->tracks.nb_elem; i++)
- if (tracks[i].num == num)
- return &tracks[i];
-
- av_log(matroska->ctx, AV_LOG_ERROR, "Invalid track number %d\n", num);
- return NULL;
+ for (i=0; syntax[i].id; i++)
+ if (id == syntax[i].id)
+ break;
+ if (!syntax[i].id && id != EBML_ID_VOID && id != EBML_ID_CRC32)
+ av_log(matroska->ctx, AV_LOG_INFO, "Unknown entry 0x%X\n", id);
+ return ebml_parse_elem(matroska, &syntax[i], data);
}
-
-/*
- * Put one packet in an application-supplied AVPacket struct.
- * Returns 0 on success or -1 on failure.
- */
-static int
-matroska_deliver_packet (MatroskaDemuxContext *matroska,
- AVPacket *pkt)
+static int ebml_parse(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
+ void *data)
{
- if (matroska->num_packets > 0) {
- memcpy(pkt, matroska->packets[0], sizeof(AVPacket));
- av_free(matroska->packets[0]);
- if (matroska->num_packets > 1) {
- memmove(&matroska->packets[0], &matroska->packets[1],
- (matroska->num_packets - 1) * sizeof(AVPacket *));
- matroska->packets =
- av_realloc(matroska->packets, (matroska->num_packets - 1) *
- sizeof(AVPacket *));
- } else {
- av_freep(&matroska->packets);
- }
- matroska->num_packets--;
- return 0;
- }
-
- return -1;
+ uint64_t id;
+ int res = ebml_read_num(matroska, matroska->ctx->pb, 4, &id);
+ id |= 1 << 7*res;
+ return res < 0 ? res : ebml_parse_id(matroska, syntax, id, data);
}
-/*
- * Put a packet into our internal queue. Will be delivered to the
- * user/application during the next get_packet() call.
- */
-static void
-matroska_queue_packet (MatroskaDemuxContext *matroska,
- AVPacket *pkt)
+static int ebml_parse_nest(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
+ void *data)
{
- matroska->packets =
- av_realloc(matroska->packets, (matroska->num_packets + 1) *
- sizeof(AVPacket *));
- matroska->packets[matroska->num_packets] = pkt;
- matroska->num_packets++;
-}
+ int i, res = 0;
-/*
- * Free all packets in our internal queue.
- */
-static void
-matroska_clear_queue (MatroskaDemuxContext *matroska)
-{
- if (matroska->packets) {
- int n;
- for (n = 0; n < matroska->num_packets; n++) {
- av_free_packet(matroska->packets[n]);
- av_free(matroska->packets[n]);
+ for (i=0; syntax[i].id; i++)
+ switch (syntax[i].type) {
+ case EBML_UINT:
+ *(uint64_t *)((char *)data+syntax[i].data_offset) = syntax[i].def.u;
+ break;
+ case EBML_FLOAT:
+ *(double *)((char *)data+syntax[i].data_offset) = syntax[i].def.f;
+ break;
+ case EBML_STR:
+ case EBML_UTF8:
+ *(char **)((char *)data+syntax[i].data_offset) = av_strdup(syntax[i].def.s);
+ break;
}
- av_free(matroska->packets);
- matroska->packets = NULL;
- matroska->num_packets = 0;
- }
-}
-
-
-/*
- * Autodetecting...
- */
-static int
-matroska_probe (AVProbeData *p)
-{
- uint64_t total = 0;
- int len_mask = 0x80, size = 1, n = 1;
- uint8_t probe_data[] = { 'm', 'a', 't', 'r', 'o', 's', 'k', 'a' };
-
- /* ebml header? */
- if (AV_RB32(p->buf) != EBML_ID_HEADER)
- return 0;
- /* length of header */
- total = p->buf[4];
- while (size <= 8 && !(total & len_mask)) {
- size++;
- len_mask >>= 1;
- }
- if (size > 8)
- return 0;
- total &= (len_mask - 1);
- while (n < size)
- total = (total << 8) | p->buf[4 + n++];
+ while (!res && !ebml_level_end(matroska))
+ res = ebml_parse(matroska, syntax, data);
- /* does the probe data contain the whole header? */
- if (p->buf_size < 4 + size + total)
- return 0;
-
- /* the header must contain the document type 'matroska'. For now,
- * we don't parse the whole header but simply check for the
- * availability of that array of characters inside the header.
- * Not fully fool-proof, but good enough. */
- for (n = 4 + size; n <= 4 + size + total - sizeof(probe_data); n++)
- if (!memcmp (&p->buf[n], probe_data, sizeof(probe_data)))
- return AVPROBE_SCORE_MAX;
-
- return 0;
+ return res;
}
-static int ebml_parse(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
- void *data, int once);
-
static int ebml_parse_elem(MatroskaDemuxContext *matroska,
EbmlSyntax *syntax, void *data)
{
}
if (syntax->type != EBML_PASS && syntax->type != EBML_STOP)
- if ((res = ebml_read_element_id(matroska, &id)) < 0 ||
- (res = ebml_read_element_length(matroska, &length)) < 0)
+ if ((res = ebml_read_num(matroska, pb, 8, &length)) < 0)
return res;
switch (syntax->type) {
return res;
if (id == MATROSKA_ID_SEGMENT)
matroska->segment_start = url_ftell(matroska->ctx->pb);
- return ebml_parse(matroska, syntax->def.n, data, 0);
- case EBML_PASS: return ebml_parse(matroska, syntax->def.n, data, 1);
+ return ebml_parse_nest(matroska, syntax->def.n, data);
+ case EBML_PASS: return ebml_parse_id(matroska, syntax->def.n, id, data);
case EBML_STOP: *(int *)data = 1; return 1;
- default: url_fskip(pb, length); return 0;
+ default: return url_fseek(pb,length,SEEK_CUR)<0 ? AVERROR(EIO) : 0;
}
if (res == AVERROR_INVALIDDATA)
av_log(matroska->ctx, AV_LOG_ERROR, "Invalid element\n");
return res;
}
-static int ebml_parse_id(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
- uint32_t id, void *data)
-{
- int i;
- for (i=0; syntax[i].id; i++)
- if (id == syntax[i].id)
- break;
- if (!syntax[i].id)
- av_log(matroska->ctx, AV_LOG_INFO, "Unknown entry 0x%X\n", id);
- return ebml_parse_elem(matroska, &syntax[i], data);
-}
-
-static int ebml_parse(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
- void *data, int once)
-{
- int i, res = 0, res2;
- uint32_t id = 0;
-
- for (i=0; syntax[i].id; i++)
- switch (syntax[i].type) {
- case EBML_UINT:
- *(uint64_t *)((char *)data+syntax[i].data_offset) = syntax[i].def.u;
- break;
- case EBML_FLOAT:
- *(double *)((char *)data+syntax[i].data_offset) = syntax[i].def.f;
- break;
- case EBML_STR:
- case EBML_UTF8:
- *(char **)((char *)data+syntax[i].data_offset) = av_strdup(syntax[i].def.s);
- break;
- }
-
- while (!res) {
- res2 = ebml_read_element_id(matroska, &id);
- if (res2 < 0)
- break;
- if (res2 > 0)
- matroska->level_up = ebml_read_element_level_up(matroska);
- if (matroska->level_up) {
- matroska->level_up--;
- break;
- }
-
- res = ebml_parse_id(matroska, syntax, id, data);
- if (once)
- break;
- }
-
- return res;
-}
-
static void ebml_free(EbmlSyntax *syntax, void *data)
{
int i, j;
}
}
-static int
-matroska_decode_buffer(uint8_t** buf, int* buf_size, MatroskaTrack *track)
+
+/*
+ * Autodetecting...
+ */
+static int matroska_probe(AVProbeData *p)
+{
+ uint64_t total = 0;
+ int len_mask = 0x80, size = 1, n = 1;
+ static const char probe_data[] = "matroska";
+
+ /* EBML header? */
+ if (AV_RB32(p->buf) != EBML_ID_HEADER)
+ return 0;
+
+ /* length of header */
+ total = p->buf[4];
+ while (size <= 8 && !(total & len_mask)) {
+ size++;
+ len_mask >>= 1;
+ }
+ if (size > 8)
+ return 0;
+ total &= (len_mask - 1);
+ while (n < size)
+ total = (total << 8) | p->buf[4 + n++];
+
+ /* Does the probe data contain the whole header? */
+ if (p->buf_size < 4 + size + total)
+ return 0;
+
+ /* The header must contain the document type 'matroska'. For now,
+ * we don't parse the whole header but simply check for the
+ * availability of that array of characters inside the header.
+ * Not fully fool-proof, but good enough. */
+ for (n = 4+size; n <= 4+size+total-(sizeof(probe_data)-1); n++)
+ if (!memcmp(p->buf+n, probe_data, sizeof(probe_data)-1))
+ return AVPROBE_SCORE_MAX;
+
+ return 0;
+}
+
+static MatroskaTrack *matroska_find_track_by_num(MatroskaDemuxContext *matroska,
+ int num)
+{
+ MatroskaTrack *tracks = matroska->tracks.elem;
+ int i;
+
+ for (i=0; i < matroska->tracks.nb_elem; i++)
+ if (tracks[i].num == num)
+ return &tracks[i];
+
+ av_log(matroska->ctx, AV_LOG_ERROR, "Invalid track number %d\n", num);
+ return NULL;
+}
+
+static int matroska_decode_buffer(uint8_t** buf, int* buf_size,
+ MatroskaTrack *track)
{
MatroskaTrackEncoding *encodings = track->encodings.elem;
uint8_t* data = *buf;
case MATROSKA_TRACK_ENCODING_COMP_LZO:
do {
olen = pkt_size *= 3;
- pkt_data = av_realloc(pkt_data,
- pkt_size+LZO_OUTPUT_PADDING);
- result = lzo1x_decode(pkt_data, &olen, data, &isize);
- } while (result==LZO_OUTPUT_FULL && pkt_size<10000000);
+ pkt_data = av_realloc(pkt_data, pkt_size+AV_LZO_OUTPUT_PADDING);
+ result = av_lzo1x_decode(pkt_data, &olen, data, &isize);
+ } while (result==AV_LZO_OUTPUT_FULL && pkt_size<10000000);
if (result)
goto failed;
pkt_size -= olen;
break;
-#ifdef CONFIG_ZLIB
+#if CONFIG_ZLIB
case MATROSKA_TRACK_ENCODING_COMP_ZLIB: {
z_stream zstream = {0};
if (inflateInit(&zstream) != Z_OK)
break;
}
#endif
-#ifdef CONFIG_BZLIB
+#if CONFIG_BZLIB
case MATROSKA_TRACK_ENCODING_COMP_BZLIB: {
bz_stream bzstream = {0};
if (BZ2_bzDecompressInit(&bzstream, 0, 0) != BZ_OK)
break;
}
#endif
+ default:
+ return -1;
}
*buf = pkt_data;
return -1;
}
-static void
-matroska_execute_seekhead(MatroskaDemuxContext *matroska)
+static void matroska_fix_ass_packet(MatroskaDemuxContext *matroska,
+ AVPacket *pkt, uint64_t display_duration)
+{
+ char *line, *layer, *ptr = pkt->data, *end = ptr+pkt->size;
+ for (; *ptr!=',' && ptr<end-1; ptr++);
+ if (*ptr == ',')
+ layer = ++ptr;
+ for (; *ptr!=',' && ptr<end-1; ptr++);
+ if (*ptr == ',') {
+ int64_t end_pts = pkt->pts + display_duration;
+ int sc = matroska->time_scale * pkt->pts / 10000000;
+ int ec = matroska->time_scale * end_pts / 10000000;
+ int sh, sm, ss, eh, em, es, len;
+ sh = sc/360000; sc -= 360000*sh;
+ sm = sc/ 6000; sc -= 6000*sm;
+ ss = sc/ 100; sc -= 100*ss;
+ eh = ec/360000; ec -= 360000*eh;
+ em = ec/ 6000; ec -= 6000*em;
+ es = ec/ 100; ec -= 100*es;
+ *ptr++ = '\0';
+ len = 50 + end-ptr + FF_INPUT_BUFFER_PADDING_SIZE;
+ if (!(line = av_malloc(len)))
+ return;
+ snprintf(line,len,"Dialogue: %s,%d:%02d:%02d.%02d,%d:%02d:%02d.%02d,%s\r\n",
+ layer, sh, sm, ss, sc, eh, em, es, ec, ptr);
+ av_free(pkt->data);
+ pkt->data = line;
+ pkt->size = strlen(line);
+ }
+}
+
+static void matroska_merge_packets(AVPacket *out, AVPacket *in)
+{
+ out->data = av_realloc(out->data, out->size+in->size);
+ memcpy(out->data+out->size, in->data, in->size);
+ out->size += in->size;
+ av_destruct_packet(in);
+ av_free(in);
+}
+
+static void matroska_convert_tag(AVFormatContext *s, EbmlList *list,
+ AVMetadata **metadata, char *prefix)
+{
+ MatroskaTag *tags = list->elem;
+ char key[1024];
+ int i;
+
+ for (i=0; i < list->nb_elem; i++) {
+ const char *lang = strcmp(tags[i].lang, "und") ? tags[i].lang : NULL;
+ if (prefix) snprintf(key, sizeof(key), "%s/%s", prefix, tags[i].name);
+ else av_strlcpy(key, tags[i].name, sizeof(key));
+ if (tags[i].def || !lang) {
+ av_metadata_set(metadata, key, tags[i].string);
+ if (tags[i].sub.nb_elem)
+ matroska_convert_tag(s, &tags[i].sub, metadata, key);
+ }
+ if (lang) {
+ av_strlcat(key, "-", sizeof(key));
+ av_strlcat(key, lang, sizeof(key));
+ av_metadata_set(metadata, key, tags[i].string);
+ if (tags[i].sub.nb_elem)
+ matroska_convert_tag(s, &tags[i].sub, metadata, key);
+ }
+ }
+}
+
+static void matroska_convert_tags(AVFormatContext *s)
+{
+ MatroskaDemuxContext *matroska = s->priv_data;
+ MatroskaTags *tags = matroska->tags.elem;
+ int i, j;
+
+ for (i=0; i < matroska->tags.nb_elem; i++) {
+ if (tags[i].target.attachuid) {
+ MatroskaAttachement *attachment = matroska->attachments.elem;
+ for (j=0; j<matroska->attachments.nb_elem; j++)
+ if (attachment[j].uid == tags[i].target.attachuid)
+ matroska_convert_tag(s, &tags[i].tag,
+ &attachment[j].stream->metadata, NULL);
+ } else if (tags[i].target.chapteruid) {
+ MatroskaChapter *chapter = matroska->chapters.elem;
+ for (j=0; j<matroska->chapters.nb_elem; j++)
+ if (chapter[j].uid == tags[i].target.chapteruid)
+ matroska_convert_tag(s, &tags[i].tag,
+ &chapter[j].chapter->metadata, NULL);
+ } else if (tags[i].target.trackuid) {
+ MatroskaTrack *track = matroska->tracks.elem;
+ for (j=0; j<matroska->tracks.nb_elem; j++)
+ if (track[j].uid == tags[i].target.trackuid)
+ matroska_convert_tag(s, &tags[i].tag,
+ &track[j].stream->metadata, NULL);
+ } else {
+ matroska_convert_tag(s, &tags[i].tag, &s->metadata,
+ tags[i].target.type);
+ }
+ }
+}
+
+static void matroska_execute_seekhead(MatroskaDemuxContext *matroska)
{
EbmlList *seekhead_list = &matroska->seekhead;
MatroskaSeekhead *seekhead = seekhead_list->elem;
- uint32_t peek_id_cache = matroska->peek_id;
uint32_t level_up = matroska->level_up;
- offset_t before_pos = url_ftell(matroska->ctx->pb);
+ int64_t before_pos = url_ftell(matroska->ctx->pb);
MatroskaLevel level;
int i;
for (i=0; i<seekhead_list->nb_elem; i++) {
+ int64_t offset = seekhead[i].pos + matroska->segment_start;
+
if (seekhead[i].pos <= before_pos
|| seekhead[i].id == MATROSKA_ID_SEEKHEAD
|| seekhead[i].id == MATROSKA_ID_CLUSTER)
continue;
/* seek */
- if (ebml_read_seek(matroska,
- seekhead[i].pos+matroska->segment_start) < 0)
+ if (url_fseek(matroska->ctx->pb, offset, SEEK_SET) != offset)
continue;
- /* we don't want to lose our seekhead level, so we add
+ /* We don't want to lose our seekhead level, so we add
* a dummy. This is a crude hack. */
if (matroska->num_levels == EBML_MAX_DEPTH) {
av_log(matroska->ctx, AV_LOG_INFO,
matroska->levels[matroska->num_levels] = level;
matroska->num_levels++;
- ebml_parse_id(matroska, matroska_segment, seekhead[i].id, matroska);
+ ebml_parse(matroska, matroska_segment, matroska);
/* remove dummy level */
while (matroska->num_levels) {
}
/* seek back */
- ebml_read_seek(matroska, before_pos);
- matroska->peek_id = peek_id_cache;
+ url_fseek(matroska->ctx->pb, before_pos, SEEK_SET);
matroska->level_up = level_up;
}
-static int
-matroska_aac_profile (char *codec_id)
+static int matroska_aac_profile(char *codec_id)
{
- static const char *aac_profiles[] = { "MAIN", "LC", "SSR" };
+ static const char * const aac_profiles[] = { "MAIN", "LC", "SSR" };
int profile;
- for (profile=0; profile<ARRAY_SIZE(aac_profiles); profile++)
+ for (profile=0; profile<FF_ARRAY_ELEMS(aac_profiles); profile++)
if (strstr(codec_id, aac_profiles[profile]))
break;
return profile + 1;
}
-static int
-matroska_aac_sri (int samplerate)
+static int matroska_aac_sri(int samplerate)
{
int sri;
- for (sri=0; sri<ARRAY_SIZE(ff_mpeg4audio_sample_rates); sri++)
+ for (sri=0; sri<FF_ARRAY_ELEMS(ff_mpeg4audio_sample_rates); sri++)
if (ff_mpeg4audio_sample_rates[sri] == samplerate)
break;
return sri;
}
-static int
-matroska_read_header (AVFormatContext *s,
- AVFormatParameters *ap)
+static int matroska_read_header(AVFormatContext *s, AVFormatParameters *ap)
{
MatroskaDemuxContext *matroska = s->priv_data;
EbmlList *attachements_list = &matroska->attachments;
MatroskaTrack *tracks;
EbmlList *index_list;
MatroskaIndex *index;
+ int index_scale = 1;
+ uint64_t max_start = 0;
Ebml ebml = { 0 };
AVStream *st;
int i, j;
matroska->ctx = s;
/* First read the EBML header. */
- if (ebml_parse(matroska, ebml_syntax, &ebml, 1)
+ if (ebml_parse(matroska, ebml_syntax, &ebml)
|| ebml.version > EBML_VERSION || ebml.max_size > sizeof(uint64_t)
|| ebml.id_length > sizeof(uint32_t) || strcmp(ebml.doctype, "matroska")
|| ebml.doctype_version > 2) {
ebml_free(ebml_syntax, &ebml);
/* The next thing is a segment. */
- if (ebml_parse(matroska, matroska_segments, matroska, 1) < 0)
+ if (ebml_parse(matroska, matroska_segments, matroska) < 0)
return -1;
matroska_execute_seekhead(matroska);
if (matroska->duration)
matroska->ctx->duration = matroska->duration * matroska->time_scale
* 1000 / AV_TIME_BASE;
- if (matroska->title)
- strncpy(matroska->ctx->title, matroska->title,
- sizeof(matroska->ctx->title)-1);
+ av_metadata_set(&s->metadata, "title", matroska->title);
tracks = matroska->tracks.elem;
for (i=0; i < matroska->tracks.nb_elem; i++) {
uint8_t *extradata = NULL;
int extradata_size = 0;
int extradata_offset = 0;
+ ByteIOContext b;
/* Apply some sanity checks. */
if (track->type != MATROSKA_TRACK_TYPE_VIDEO &&
} else if (encodings_list->nb_elem == 1) {
if (encodings[0].type ||
(encodings[0].compression.algo != MATROSKA_TRACK_ENCODING_COMP_HEADERSTRIP &&
-#ifdef CONFIG_ZLIB
+#if CONFIG_ZLIB
encodings[0].compression.algo != MATROSKA_TRACK_ENCODING_COMP_ZLIB &&
#endif
-#ifdef CONFIG_BZLIB
+#if CONFIG_BZLIB
encodings[0].compression.algo != MATROSKA_TRACK_ENCODING_COMP_BZLIB &&
#endif
encodings[0].compression.algo != MATROSKA_TRACK_ENCODING_COMP_LZO)) {
}
}
- st = track->stream = av_new_stream(s, matroska->num_streams++);
+ st = track->stream = av_new_stream(s, 0);
if (st == NULL)
return AVERROR(ENOMEM);
- if (!strcmp(track->codec_id, MATROSKA_CODEC_ID_VIDEO_VFW_FOURCC)
+ if (!strcmp(track->codec_id, "V_MS/VFW/FOURCC")
&& track->codec_priv.size >= 40
&& track->codec_priv.data != NULL) {
+ track->ms_compat = 1;
track->video.fourcc = AV_RL32(track->codec_priv.data + 16);
- codec_id = codec_get_id(codec_bmp_tags, track->video.fourcc);
- } else if (!strcmp(track->codec_id, MATROSKA_CODEC_ID_AUDIO_ACM)
- && track->codec_priv.size >= 18
+ codec_id = ff_codec_get_id(ff_codec_bmp_tags, track->video.fourcc);
+ extradata_offset = 40;
+ } else if (!strcmp(track->codec_id, "A_MS/ACM")
+ && track->codec_priv.size >= 14
&& track->codec_priv.data != NULL) {
- uint16_t tag = AV_RL16(track->codec_priv.data);
- codec_id = codec_get_id(codec_wav_tags, tag);
+ init_put_byte(&b, track->codec_priv.data, track->codec_priv.size,
+ URL_RDONLY, NULL, NULL, NULL, NULL);
+ ff_get_wav_header(&b, st->codec, track->codec_priv.size);
+ codec_id = st->codec->codec_id;
+ extradata_offset = FFMIN(track->codec_priv.size, 18);
} else if (!strcmp(track->codec_id, "V_QUICKTIME")
&& (track->codec_priv.size >= 86)
&& (track->codec_priv.data != NULL)) {
track->video.fourcc = AV_RL32(track->codec_priv.data);
- codec_id=codec_get_id(codec_movvideo_tags, track->video.fourcc);
+ codec_id=ff_codec_get_id(codec_movvideo_tags, track->video.fourcc);
+ } else if (codec_id == CODEC_ID_PCM_S16BE) {
+ switch (track->audio.bitdepth) {
+ case 8: codec_id = CODEC_ID_PCM_U8; break;
+ case 24: codec_id = CODEC_ID_PCM_S24BE; break;
+ case 32: codec_id = CODEC_ID_PCM_S32BE; break;
+ }
+ } else if (codec_id == CODEC_ID_PCM_S16LE) {
+ switch (track->audio.bitdepth) {
+ case 8: codec_id = CODEC_ID_PCM_U8; break;
+ case 24: codec_id = CODEC_ID_PCM_S24LE; break;
+ case 32: codec_id = CODEC_ID_PCM_S32LE; break;
+ }
+ } else if (codec_id==CODEC_ID_PCM_F32LE && track->audio.bitdepth==64) {
+ codec_id = CODEC_ID_PCM_F64LE;
} else if (codec_id == CODEC_ID_AAC && !track->codec_priv.size) {
int profile = matroska_aac_profile(track->codec_id);
int sri = matroska_aac_sri(track->audio.samplerate);
} else
extradata_size = 2;
} else if (codec_id == CODEC_ID_TTA) {
- ByteIOContext b;
extradata_size = 30;
extradata = av_mallocz(extradata_size);
if (extradata == NULL)
} else if (codec_id == CODEC_ID_RV10 || codec_id == CODEC_ID_RV20 ||
codec_id == CODEC_ID_RV30 || codec_id == CODEC_ID_RV40) {
extradata_offset = 26;
- track->codec_priv.size -= extradata_offset;
} else if (codec_id == CODEC_ID_RA_144) {
track->audio.out_samplerate = 8000;
track->audio.channels = 1;
} else if (codec_id == CODEC_ID_RA_288 || codec_id == CODEC_ID_COOK ||
codec_id == CODEC_ID_ATRAC3) {
- ByteIOContext b;
-
init_put_byte(&b, track->codec_priv.data,track->codec_priv.size,
0, NULL, NULL, NULL, NULL);
url_fskip(&b, 24);
} else {
st->codec->block_align = track->audio.sub_packet_size;
extradata_offset = 78;
- track->codec_priv.size -= extradata_offset;
}
}
+ track->codec_priv.size -= extradata_offset;
if (codec_id == CODEC_ID_NONE)
av_log(matroska->ctx, AV_LOG_INFO,
"Unknown/unsupported CodecID %s.\n", track->codec_id);
+ if (track->time_scale < 0.01)
+ track->time_scale = 1.0;
av_set_pts_info(st, 64, matroska->time_scale*track->time_scale, 1000*1000*1000); /* 64 bit pts in ns */
st->codec->codec_id = codec_id;
st->start_time = 0;
if (strcmp(track->language, "und"))
- av_strlcpy(st->language, track->language, 4);
+ av_metadata_set(&st->metadata, "language", track->language);
+ av_metadata_set(&st->metadata, "title", track->name);
if (track->flag_default)
st->disposition |= AV_DISPOSITION_DEFAULT;
av_reduce(&st->codec->time_base.num, &st->codec->time_base.den,
track->default_duration, 1000000000, 30000);
- if(extradata){
- st->codec->extradata = extradata;
- st->codec->extradata_size = extradata_size;
- } else if(track->codec_priv.data && track->codec_priv.size > 0){
- st->codec->extradata = av_malloc(track->codec_priv.size);
- if(st->codec->extradata == NULL)
- return AVERROR(ENOMEM);
- st->codec->extradata_size = track->codec_priv.size;
- memcpy(st->codec->extradata,
- track->codec_priv.data + extradata_offset,
- track->codec_priv.size);
+ if (!st->codec->extradata) {
+ if(extradata){
+ st->codec->extradata = extradata;
+ st->codec->extradata_size = extradata_size;
+ } else if(track->codec_priv.data && track->codec_priv.size > 0){
+ st->codec->extradata = av_mallocz(track->codec_priv.size +
+ FF_INPUT_BUFFER_PADDING_SIZE);
+ if(st->codec->extradata == NULL)
+ return AVERROR(ENOMEM);
+ st->codec->extradata_size = track->codec_priv.size;
+ memcpy(st->codec->extradata,
+ track->codec_priv.data + extradata_offset,
+ track->codec_priv.size);
+ }
}
if (track->type == MATROSKA_TRACK_TYPE_VIDEO) {
st->codec->codec_tag = track->video.fourcc;
st->codec->width = track->video.pixel_width;
st->codec->height = track->video.pixel_height;
- av_reduce(&st->codec->sample_aspect_ratio.num,
- &st->codec->sample_aspect_ratio.den,
+ av_reduce(&st->sample_aspect_ratio.num,
+ &st->sample_aspect_ratio.den,
st->codec->height * track->video.display_width,
st->codec-> width * track->video.display_height,
255);
+ if (st->codec->codec_id != CODEC_ID_H264)
st->need_parsing = AVSTREAM_PARSE_HEADERS;
} else if (track->type == MATROSKA_TRACK_TYPE_AUDIO) {
st->codec->codec_type = CODEC_TYPE_AUDIO;
attachements[j].bin.data && attachements[j].bin.size > 0)) {
av_log(matroska->ctx, AV_LOG_ERROR, "incomplete attachment\n");
} else {
- AVStream *st = av_new_stream(s, matroska->num_streams++);
+ AVStream *st = av_new_stream(s, 0);
if (st == NULL)
break;
- st->filename = av_strdup(attachements[j].filename);
+ av_metadata_set(&st->metadata, "filename",attachements[j].filename);
st->codec->codec_id = CODEC_ID_NONE;
st->codec->codec_type = CODEC_TYPE_ATTACHMENT;
st->codec->extradata = av_malloc(attachements[j].bin.size);
break;
}
}
+ attachements[j].stream = st;
}
}
chapters = chapters_list->elem;
for (i=0; i<chapters_list->nb_elem; i++)
- if (chapters[i].start != AV_NOPTS_VALUE && chapters[i].uid)
+ if (chapters[i].start != AV_NOPTS_VALUE && chapters[i].uid
+ && (max_start==0 || chapters[i].start > max_start)) {
+ chapters[i].chapter =
ff_new_chapter(s, chapters[i].uid, (AVRational){1, 1000000000},
chapters[i].start, chapters[i].end,
chapters[i].title);
+ av_metadata_set(&chapters[i].chapter->metadata,
+ "title", chapters[i].title);
+ max_start = chapters[i].start;
+ }
index_list = &matroska->index;
index = index_list->elem;
+ if (index_list->nb_elem
+ && index[0].time > 100000000000000/matroska->time_scale) {
+ av_log(matroska->ctx, AV_LOG_WARNING, "Working around broken index.\n");
+ index_scale = matroska->time_scale;
+ }
for (i=0; i<index_list->nb_elem; i++) {
EbmlList *pos_list = &index[i].pos;
MatroskaIndexPos *pos = pos_list->elem;
if (track && track->stream)
av_add_index_entry(track->stream,
pos[j].pos + matroska->segment_start,
- index[i].time*matroska->time_scale/AV_TIME_BASE,
- 0, 0, AVINDEX_KEYFRAME);
+ index[i].time/index_scale, 0, 0,
+ AVINDEX_KEYFRAME);
}
}
+ matroska_convert_tags(s);
+
return 0;
}
-static int
-matroska_parse_block(MatroskaDemuxContext *matroska, uint8_t *data, int size,
- int64_t pos, uint64_t cluster_time, uint64_t duration,
- int is_keyframe)
+/*
+ * Put one packet in an application-supplied AVPacket struct.
+ * Returns 0 on success or -1 on failure.
+ */
+static int matroska_deliver_packet(MatroskaDemuxContext *matroska,
+ AVPacket *pkt)
{
+ if (matroska->num_packets > 0) {
+ memcpy(pkt, matroska->packets[0], sizeof(AVPacket));
+ av_free(matroska->packets[0]);
+ if (matroska->num_packets > 1) {
+ memmove(&matroska->packets[0], &matroska->packets[1],
+ (matroska->num_packets - 1) * sizeof(AVPacket *));
+ matroska->packets =
+ av_realloc(matroska->packets, (matroska->num_packets - 1) *
+ sizeof(AVPacket *));
+ } else {
+ av_freep(&matroska->packets);
+ }
+ matroska->num_packets--;
+ return 0;
+ }
+
+ return -1;
+}
+
+/*
+ * Free all packets in our internal queue.
+ */
+static void matroska_clear_queue(MatroskaDemuxContext *matroska)
+{
+ if (matroska->packets) {
+ int n;
+ for (n = 0; n < matroska->num_packets; n++) {
+ av_free_packet(matroska->packets[n]);
+ av_free(matroska->packets[n]);
+ }
+ av_freep(&matroska->packets);
+ matroska->num_packets = 0;
+ }
+}
+
+static int matroska_parse_block(MatroskaDemuxContext *matroska, uint8_t *data,
+ int size, int64_t pos, uint64_t cluster_time,
+ uint64_t duration, int is_keyframe,
+ int64_t cluster_pos)
+{
+ uint64_t timecode = AV_NOPTS_VALUE;
MatroskaTrack *track;
int res = 0;
AVStream *st;
int n, flags, laces = 0;
uint64_t num;
- if ((n = matroska_ebmlnum_uint(data, size, &num)) < 0) {
+ if ((n = matroska_ebmlnum_uint(matroska, data, size, &num)) < 0) {
av_log(matroska->ctx, AV_LOG_ERROR, "EBML block data error\n");
return res;
}
if (is_keyframe == -1)
is_keyframe = flags & 0x80 ? PKT_FLAG_KEY : 0;
- if (matroska->skip_to_keyframe) {
- if (!is_keyframe || st != matroska->skip_to_stream)
+ if (cluster_time != (uint64_t)-1
+ && (block_time >= 0 || cluster_time >= -block_time)) {
+ timecode = cluster_time + block_time;
+ if (track->type == MATROSKA_TRACK_TYPE_SUBTITLE
+ && timecode < track->end_timecode)
+ is_keyframe = 0; /* overlapping subtitles are not key frame */
+ if (is_keyframe)
+ av_add_index_entry(st, cluster_pos, timecode, 0,0,AVINDEX_KEYFRAME);
+ track->end_timecode = FFMAX(track->end_timecode, timecode+duration);
+ }
+
+ if (matroska->skip_to_keyframe && track->type != MATROSKA_TRACK_TYPE_SUBTITLE) {
+ if (!is_keyframe || timecode < matroska->skip_to_timecode)
return res;
matroska->skip_to_keyframe = 0;
}
lace_size[0] = size;
break;
- case 0x1: /* xiph lacing */
+ case 0x1: /* Xiph lacing */
case 0x2: /* fixed-size lacing */
case 0x3: /* EBML lacing */
assert(size>0); // size <=3 is checked before size-=3 above
lace_size = av_mallocz(laces * sizeof(int));
switch ((flags & 0x06) >> 1) {
- case 0x1: /* xiph lacing */ {
+ case 0x1: /* Xiph lacing */ {
uint8_t temp;
uint32_t total = 0;
for (n = 0; res == 0 && n < laces - 1; n++) {
case 0x3: /* EBML lacing */ {
uint32_t total;
- n = matroska_ebmlnum_uint(data, size, &num);
+ n = matroska_ebmlnum_uint(matroska, data, size, &num);
if (n < 0) {
av_log(matroska->ctx, AV_LOG_INFO,
"EBML block data error\n");
for (n = 1; res == 0 && n < laces - 1; n++) {
int64_t snum;
int r;
- r = matroska_ebmlnum_sint (data, size, &snum);
+ r = matroska_ebmlnum_sint(matroska, data, size, &snum);
if (r < 0) {
av_log(matroska->ctx, AV_LOG_INFO,
"EBML block data error\n");
}
if (res == 0) {
- uint64_t timecode = AV_NOPTS_VALUE;
-
- if (cluster_time != (uint64_t)-1
- && (block_time >= 0 || cluster_time >= -block_time))
- timecode = cluster_time + block_time;
-
for (n = 0; n < laces; n++) {
- if (st->codec->codec_id == CODEC_ID_RA_288 ||
- st->codec->codec_id == CODEC_ID_COOK ||
- st->codec->codec_id == CODEC_ID_ATRAC3) {
+ if ((st->codec->codec_id == CODEC_ID_RA_288 ||
+ st->codec->codec_id == CODEC_ID_COOK ||
+ st->codec->codec_id == CODEC_ID_ATRAC3) &&
+ st->codec->block_align && track->audio.sub_packet_size) {
int a = st->codec->block_align;
int sps = track->audio.sub_packet_size;
int cfs = track->audio.coded_framesize;
+ a * (h*w / a - track->audio.pkt_cnt--), a);
pkt->pos = pos;
pkt->stream_index = st->index;
- matroska_queue_packet(matroska, pkt);
+ dynarray_add(&matroska->packets,&matroska->num_packets,pkt);
}
} else {
MatroskaTrackEncoding *encodings = track->encodings.elem;
int offset = 0, pkt_size = lace_size[n];
uint8_t *pkt_data = data;
+ if (lace_size[n] > size) {
+ av_log(matroska->ctx, AV_LOG_ERROR, "Invalid packet size\n");
+ break;
+ }
+
if (encodings && encodings->scope & 1) {
offset = matroska_decode_buffer(&pkt_data,&pkt_size, track);
if (offset < 0)
if (av_new_packet(pkt, pkt_size+offset) < 0) {
av_free(pkt);
res = AVERROR(ENOMEM);
- n = laces-1;
break;
}
if (offset)
pkt->flags = is_keyframe;
pkt->stream_index = st->index;
- pkt->pts = timecode;
+ if (track->ms_compat)
+ pkt->dts = timecode;
+ else
+ pkt->pts = timecode;
pkt->pos = pos;
- pkt->duration = duration;
-
- matroska_queue_packet(matroska, pkt);
+ if (st->codec->codec_id == CODEC_ID_TEXT)
+ pkt->convergence_duration = duration;
+ else if (track->type != MATROSKA_TRACK_TYPE_SUBTITLE)
+ pkt->duration = duration;
+
+ if (st->codec->codec_id == CODEC_ID_SSA)
+ matroska_fix_ass_packet(matroska, pkt, duration);
+
+ if (matroska->prev_pkt &&
+ timecode != AV_NOPTS_VALUE &&
+ matroska->prev_pkt->pts == timecode &&
+ matroska->prev_pkt->stream_index == st->index)
+ matroska_merge_packets(matroska->prev_pkt, pkt);
+ else {
+ dynarray_add(&matroska->packets,&matroska->num_packets,pkt);
+ matroska->prev_pkt = pkt;
+ }
}
if (timecode != AV_NOPTS_VALUE)
timecode = duration ? timecode + duration : AV_NOPTS_VALUE;
data += lace_size[n];
+ size -= lace_size[n];
}
}
return res;
}
-static int
-matroska_parse_cluster (MatroskaDemuxContext *matroska)
+static int matroska_parse_cluster(MatroskaDemuxContext *matroska)
{
MatroskaCluster cluster = { 0 };
EbmlList *blocks_list;
MatroskaBlock *blocks;
- int i, res = ebml_parse(matroska, matroska_clusters, &cluster, 1);
+ int i, res;
+ int64_t pos = url_ftell(matroska->ctx->pb);
+ matroska->prev_pkt = NULL;
+ if (matroska->has_cluster_id){
+ /* For the first cluster we parse, its ID was already read as
+ part of matroska_read_header(), so don't read it again */
+ res = ebml_parse_id(matroska, matroska_clusters,
+ MATROSKA_ID_CLUSTER, &cluster);
+ pos -= 4; /* sizeof the ID which was already read */
+ matroska->has_cluster_id = 0;
+ } else
+ res = ebml_parse(matroska, matroska_clusters, &cluster);
blocks_list = &cluster.blocks;
blocks = blocks_list->elem;
- for (i=0; !res && i<blocks_list->nb_elem; i++)
- if (blocks[i].bin.size > 0)
+ for (i=0; i<blocks_list->nb_elem; i++)
+ if (blocks[i].bin.size > 0) {
+ int is_keyframe = blocks[i].non_simple ? !blocks[i].reference : -1;
res=matroska_parse_block(matroska,
blocks[i].bin.data, blocks[i].bin.size,
blocks[i].bin.pos, cluster.timecode,
- blocks[i].duration, !blocks[i].reference);
+ blocks[i].duration, is_keyframe,
+ pos);
+ }
ebml_free(matroska_cluster, &cluster);
+ if (res < 0) matroska->done = 1;
return res;
}
-static int
-matroska_read_packet (AVFormatContext *s,
- AVPacket *pkt)
+static int matroska_read_packet(AVFormatContext *s, AVPacket *pkt)
{
MatroskaDemuxContext *matroska = s->priv_data;
while (matroska_deliver_packet(matroska, pkt)) {
if (matroska->done)
- return AVERROR(EIO);
- if (matroska_parse_cluster(matroska) < 0)
- matroska->done = 1;
+ return AVERROR_EOF;
+ matroska_parse_cluster(matroska);
}
return 0;
}
-static int
-matroska_read_seek (AVFormatContext *s, int stream_index, int64_t timestamp,
- int flags)
+static int matroska_read_seek(AVFormatContext *s, int stream_index,
+ int64_t timestamp, int flags)
{
MatroskaDemuxContext *matroska = s->priv_data;
+ MatroskaTrack *tracks = matroska->tracks.elem;
AVStream *st = s->streams[stream_index];
- int index;
+ int i, index, index_sub, index_min;
- index = av_index_search_timestamp(st, timestamp, flags);
- if (index < 0)
+ if (!st->nb_index_entries)
return 0;
+ timestamp = FFMAX(timestamp, st->index_entries[0].timestamp);
+
+ if ((index = av_index_search_timestamp(st, timestamp, flags)) < 0) {
+ url_fseek(s->pb, st->index_entries[st->nb_index_entries-1].pos, SEEK_SET);
+ while ((index = av_index_search_timestamp(st, timestamp, flags)) < 0) {
+ matroska_clear_queue(matroska);
+ if (matroska_parse_cluster(matroska) < 0)
+ break;
+ }
+ }
matroska_clear_queue(matroska);
+ if (index < 0)
+ return 0;
+
+ index_min = index;
+ for (i=0; i < matroska->tracks.nb_elem; i++) {
+ tracks[i].end_timecode = 0;
+ if (tracks[i].type == MATROSKA_TRACK_TYPE_SUBTITLE
+ && !tracks[i].stream->discard != AVDISCARD_ALL) {
+ index_sub = av_index_search_timestamp(tracks[i].stream, st->index_entries[index].timestamp, AVSEEK_FLAG_BACKWARD);
+ if (index_sub >= 0
+ && st->index_entries[index_sub].pos < st->index_entries[index_min].pos
+ && st->index_entries[index].timestamp - st->index_entries[index_sub].timestamp < 30000000000/matroska->time_scale)
+ index_min = index_sub;
+ }
+ }
- url_fseek(s->pb, st->index_entries[index].pos, SEEK_SET);
+ url_fseek(s->pb, st->index_entries[index_min].pos, SEEK_SET);
matroska->skip_to_keyframe = !(flags & AVSEEK_FLAG_ANY);
- matroska->skip_to_stream = st;
- matroska->peek_id = 0;
+ matroska->skip_to_timecode = st->index_entries[index].timestamp;
+ matroska->done = 0;
av_update_cur_dts(s, st, st->index_entries[index].timestamp);
return 0;
}
-static int
-matroska_read_close (AVFormatContext *s)
+static int matroska_read_close(AVFormatContext *s)
{
MatroskaDemuxContext *matroska = s->priv_data;
MatroskaTrack *tracks = matroska->tracks.elem;
matroska_read_packet,
matroska_read_close,
matroska_read_seek,
+ .metadata_conv = ff_mkv_metadata_conv,
};
projects / ffmpeg / blobdiff