/*
- * 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.
*
* 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 "avformat.h"
EbmlList pos;
} MatroskaIndex;
+typedef struct {
+ char *name;
+ char *string;
+ EbmlList sub;
+} MatroskaTag;
+
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;
- /* The packet queue. */
+ /* the packet queue */
AVPacket **packets;
int num_packets;
{ 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 }
};
{ 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 }
};
{ 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_FILEDESC, EBML_NONE },
{ MATROSKA_ID_FILEUID, EBML_NONE },
- { EBML_ID_VOID, 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_SIMPLETAG, EBML_NEST, sizeof(MatroskaTag), offsetof(MatroskaTag,sub), {.n=matroska_simpletag} },
+ { MATROSKA_ID_TAGLANG, EBML_NONE },
+ { MATROSKA_ID_TAGDEFAULT, EBML_NONE },
+ { 0 }
+};
+
+static EbmlSyntax matroska_tag[] = {
+ { MATROSKA_ID_SIMPLETAG, EBML_NEST, sizeof(MatroskaTag), 0, {.n=matroska_simpletag} },
+ { MATROSKA_ID_TAGTARGETS, EBML_NONE },
{ 0 }
};
static EbmlSyntax matroska_tags[] = {
- { EBML_ID_VOID, EBML_NONE },
+ { MATROSKA_ID_TAG, EBML_NEST, 0, 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 },
{ 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 }
};
+#define SIZE_OFF(x) sizeof(((AVFormatContext*)0)->x),offsetof(AVFormatContext,x)
+const struct {
+ const char name[16];
+ int size;
+ int offset;
+} metadata[] = {
+ { "TITLE", SIZE_OFF(title) },
+ { "ARTIST", SIZE_OFF(author) },
+ { "WRITTEN_BY", SIZE_OFF(author) },
+ { "LEAD_PERFORMER", SIZE_OFF(author) },
+ { "COPYRIGHT", SIZE_OFF(copyright) },
+ { "COMMENT", SIZE_OFF(comment) },
+ { "ALBUM", SIZE_OFF(album) },
+ { "DATE_WRITTEN", SIZE_OFF(year) },
+ { "DATE_RELEASED", SIZE_OFF(year) },
+ { "PART_NUMBER", SIZE_OFF(track) },
+ { "GENRE", SIZE_OFF(genre) },
+};
+
/*
- * Return: whether we reached the end of a level in the hierarchy or not
+ * Return: Whether we reached the end of a level in the hierarchy or not.
*/
static int ebml_level_end(MatroskaDemuxContext *matroska)
{
* 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,
+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))) {
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.
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);
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.
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;
- }
-}
+ while (!res && !ebml_level_end(matroska))
+ res = ebml_parse(matroska, syntax, data);
-/*
- * Autodetecting...
- */
-static int matroska_probe(AVProbeData *p)
-{
- uint64_t total = 0;
- int len_mask = 0x80, size = 1, n = 1;
- 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;
+ 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 && !ebml_level_end(matroska)) {
- res2 = ebml_read_element_id(matroska, &id);
- if (res2 < 0)
- 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;
}
}
+
+/*
+ * Autodetecting...
+ */
+static int matroska_probe(AVProbeData *p)
+{
+ uint64_t total = 0;
+ int len_mask = 0x80, size = 1, n = 1;
+ 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)
{
break;
}
#endif
+ default:
+ return -1;
}
*buf = pkt_data;
return -1;
}
+static void matroska_convert_tags(AVFormatContext *s, EbmlList *list)
+{
+ MatroskaTag *tags = list->elem;
+ int i, j;
+
+ for (i=0; i < list->nb_elem; i++) {
+ for (j=0; j < ARRAY_SIZE(metadata); j++){
+ if (!strcmp(tags[i].name, metadata[j].name)) {
+ int *ptr = (int *)((char *)s + metadata[j].offset);
+ if (*ptr) continue;
+ if (metadata[j].size > sizeof(int))
+ av_strlcpy((char *)ptr, tags[i].string, metadata[j].size);
+ else
+ *ptr = atoi(tags[i].string);
+ }
+ }
+ if (tags[i].sub.nb_elem)
+ matroska_convert_tags(s, &tags[i].sub);
+ }
+}
+
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);
MatroskaLevel level;
int i;
for (i=0; i<seekhead_list->nb_elem; i++) {
+ offset_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;
}
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->title)
strncpy(matroska->ctx->title, matroska->title,
sizeof(matroska->ctx->title)-1);
+ matroska_convert_tags(s, &matroska->tags);
tracks = matroska->tracks.elem;
for (i=0; i < matroska->tracks.nb_elem; i++) {
}
}
- 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->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)
+ } else if (!strcmp(track->codec_id, "A_MS/ACM")
&& track->codec_priv.size >= 18
&& track->codec_priv.data != NULL) {
uint16_t tag = AV_RL16(track->codec_priv.data);
&& (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);
+ } 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);
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);
return 0;
}
+/*
+ * 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)
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;
}
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");
+ 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;
pkt->pos = pos;
pkt->duration = duration;
- matroska_queue_packet(matroska, pkt);
+ dynarray_add(&matroska->packets, &matroska->num_packets, pkt);
}
if (timecode != AV_NOPTS_VALUE)
MatroskaCluster cluster = { 0 };
EbmlList *blocks_list;
MatroskaBlock *blocks;
- int i, res = ebml_parse(matroska, matroska_clusters, &cluster, 1);
+ int i, res;
+ 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);
+ 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++)
+ for (i=0; i<blocks_list->nb_elem; i++)
if (blocks[i].bin.size > 0)
res=matroska_parse_block(matroska,
blocks[i].bin.data, blocks[i].bin.size,
url_fseek(s->pb, st->index_entries[index].pos, SEEK_SET);
matroska->skip_to_keyframe = !(flags & AVSEEK_FLAG_ANY);
matroska->skip_to_stream = st;
- matroska->peek_id = 0;
av_update_cur_dts(s, st, st->index_entries[index].timestamp);
return 0;
}