X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavformat%2Fmov.c;h=76c25a3e4d6de9d22f1d54d7f99f8bbabb3a0762;hb=3438d82d4b3bd987304975961e2a42e82767107d;hp=5bcd953fcd7fdeb6e9fb83eaa42d3ba0982b39ac;hpb=03dc32f689efaac92f2824b823782ecd0b480151;p=ffmpeg diff --git a/libavformat/mov.c b/libavformat/mov.c index 5bcd953fcd7..76c25a3e4d6 100644 --- a/libavformat/mov.c +++ b/libavformat/mov.c @@ -27,6 +27,8 @@ #include "riff.h" #include "isom.h" #include "dv.h" +#include "libavcodec/mpeg4audio.h" +#include "libavcodec/mpegaudiodata.h" #ifdef CONFIG_ZLIB #include @@ -69,20 +71,39 @@ typedef struct { int id; } MOV_stsc_t; +typedef struct { + uint32_t type; + char *path; +} MOV_dref_t; + typedef struct { uint32_t type; int64_t offset; int64_t size; /* total size (excluding the size and type fields) */ } MOV_atom_t; +struct MOVParseTableEntry; + typedef struct { - offset_t offset; - int64_t size; -} MOV_mdat_t; + unsigned track_id; + uint64_t base_data_offset; + uint64_t moof_offset; + unsigned stsd_id; + unsigned duration; + unsigned size; + unsigned flags; +} MOVFragment; -struct MOVParseTableEntry; +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; @@ -107,6 +128,11 @@ typedef struct MOVStreamContext { 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; + MOV_dref_t *drefs; + int dref_id; } MOVStreamContext; typedef struct MOVContext { @@ -115,15 +141,13 @@ typedef struct MOVContext { 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 */ - int64_t mdat_offset; - int total_streams; - MOVStreamContext *streams[MAX_STREAMS]; AVPaletteControl palette_control; - MOV_mdat_t *mdat_list; - int mdat_count; 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; } MOVContext; @@ -131,10 +155,9 @@ typedef struct MOVContext { /* those functions parse an atom */ /* return code: - 1: found what I wanted, exit - 0: continue to parse next atom - -1: error occured, exit - */ + 0: continue to parse next atom + <0: error occurred, exit +*/ /* links atom IDs to parse functions */ typedef struct MOVParseTableEntry { uint32_t type; @@ -156,14 +179,15 @@ static int mov_read_default(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) atom.size = INT64_MAX; while(((total_size + 8) < atom.size) && !url_feof(pb) && !err) { a.size = atom.size; - a.type=0L; + a.type=0; if(atom.size >= 8) { a.size = get_be32(pb); a.type = get_le32(pb); } total_size += 8; a.offset += 8; - dprintf(c->fc, "type: %08x %.4s sz: %"PRIx64" %"PRIx64" %"PRIx64"\n", a.type, (char*)&a.type, a.size, atom.size, total_size); + dprintf(c->fc, "type: %08x %.4s sz: %"PRIx64" %"PRIx64" %"PRIx64"\n", + a.type, (char*)&a.type, a.size, atom.size, total_size); if (a.size == 1) { /* 64 bit extended size */ a.size = get_be64(pb) - 8; a.offset += 8; @@ -179,7 +203,7 @@ static int mov_read_default(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) break; a.size = FFMIN(a.size, atom.size - total_size); - for (i = 0; mov_default_parse_table[i].type != 0L + for (i = 0; mov_default_parse_table[i].type != 0 && mov_default_parse_table[i].type != a.type; i++) /* empty */; @@ -189,7 +213,7 @@ static int mov_read_default(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) offset_t start_pos = url_ftell(pb); int64_t left; err = mov_default_parse_table[i].parse(c, pb, a); - if (c->found_moov && c->found_mdat) + if (url_is_streamed(pb) && c->found_moov && c->found_mdat) break; left = a.size - url_ftell(pb) + start_pos; if (left > 0) /* skip garbage at atom end */ @@ -200,13 +224,80 @@ static int mov_read_default(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) total_size += a.size; } - if (!err && total_size < atom.size && atom.size < 0x7ffff) { + if (!err && total_size < atom.size && atom.size < 0x7ffff) url_fskip(pb, atom.size - total_size); - } return err; } +static int mov_read_dref(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) +{ + AVStream *st = c->fc->streams[c->fc->nb_streams-1]; + MOVStreamContext *sc = st->priv_data; + int entries, i, j; + + get_be32(pb); // version + flags + 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)); + + for (i = 0; i < sc->drefs_count; i++) { + MOV_dref_t *dref = &sc->drefs[i]; + uint32_t size = get_be32(pb); + offset_t next = url_ftell(pb) + size - 4; + + dref->type = get_le32(pb); + get_be32(pb); // version + flags + dprintf(c->fc, "type %.4s size %d\n", (char*)&dref->type, size); + + if (dref->type == MKTAG('a','l','i','s') && size > 150) { + /* macintosh alias record */ + uint16_t volume_len, len; + char volume[28]; + int16_t type; + + url_fskip(pb, 10); + + volume_len = get_byte(pb); + volume_len = FFMIN(volume_len, 27); + get_buffer(pb, volume, 27); + volume[volume_len] = 0; + av_log(c->fc, AV_LOG_DEBUG, "volume %s, len %d\n", volume, volume_len); + + url_fskip(pb, 112); + + for (type = 0; type != -1 && url_ftell(pb) < next; ) { + type = get_be16(pb); + len = get_be16(pb); + av_log(c->fc, AV_LOG_DEBUG, "type %d, len %d\n", type, len); + if (len&1) + len += 1; + if (type == 2) { // absolute path + av_free(dref->path); + dref->path = av_mallocz(len+1); + if (!dref->path) + return AVERROR(ENOMEM); + get_buffer(pb, dref->path, len); + if (len > volume_len && !strncmp(dref->path, volume, volume_len)) { + len -= volume_len; + memmove(dref->path, dref->path+volume_len, len); + dref->path[len] = 0; + } + for (j = 0; j < len; j++) + if (dref->path[j] == ':') + dref->path[j] = '/'; + av_log(c->fc, AV_LOG_DEBUG, "path %s\n", dref->path); + } else + url_fskip(pb, len); + } + } + url_fseek(pb, next, SEEK_SET); + } + return 0; +} + static int mov_read_hdlr(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { AVStream *st = c->fc->streams[c->fc->nb_streams-1]; @@ -214,25 +305,26 @@ static int mov_read_hdlr(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) uint32_t ctype; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ /* component type */ 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]); + 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; - if(type == MKTAG('v', 'i', 'd', 'e')) + if (type == MKTAG('v','i','d','e')) st->codec->codec_type = CODEC_TYPE_VIDEO; - else if(type == MKTAG('s', 'o', 'u', 'n')) + else if(type == MKTAG('s','o','u','n')) st->codec->codec_type = CODEC_TYPE_AUDIO; - else if(type == MKTAG('m', '1', 'a', ' ')) + 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; - st->codec->codec_id = CODEC_ID_DVD_SUBTITLE; } get_be32(pb); /* component manufacture */ get_be32(pb); /* component flags */ @@ -271,6 +363,14 @@ static int mp4_read_descr(MOVContext *c, ByteIOContext *pb, int *tag) #define MP4DecConfigDescrTag 0x04 #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 }, +}; + static int mov_read_esds(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { AVStream *st = c->fc->streams[c->fc->nb_streams-1]; @@ -297,14 +397,31 @@ static int mov_read_esds(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) len = mp4_read_descr(c, pb, &tag); if (tag == MP4DecSpecificDescrTag) { dprintf(c->fc, "Specific MPEG4 header len=%d\n", len); + if((uint64_t)len > (1<<30)) + return -1; st->codec->extradata = av_mallocz(len + FF_INPUT_BUFFER_PADDING_SIZE); - if (st->codec->extradata) { - get_buffer(pb, st->codec->extradata, len); - st->codec->extradata_size = len; - /* from mplayer */ - if ((*st->codec->extradata >> 3) == 29) { - st->codec->codec_id = CODEC_ID_MP3ON4; - } + if (!st->codec->extradata) + return AVERROR(ENOMEM); + get_buffer(pb, st->codec->extradata, len); + st->codec->extradata_size = len; + if (st->codec->codec_id == CODEC_ID_AAC) { + MPEG4AudioConfig cfg; + ff_mpeg4audio_get_config(&cfg, st->codec->extradata, + st->codec->extradata_size); + if (cfg.chan_config > 7) + return -1; + st->codec->channels = ff_mpeg4audio_channels[cfg.chan_config]; + if (cfg.object_type == 29 && cfg.sampling_index < 3) // old mp3on4 + st->codec->sample_rate = ff_mpa_freq_tab[cfg.sampling_index]; + else + st->codec->sample_rate = cfg.sample_rate; // ext sample rate ? + dprintf(c->fc, "mp4a config channels %d obj %d ext obj %d " + "sample rate %d ext sample rate %d\n", st->codec->channels, + cfg.object_type, cfg.ext_object_type, + cfg.sample_rate, cfg.ext_sample_rate); + if (!(st->codec->codec_id = codec_get_id(mp4_audio_types, + cfg.object_type))) + st->codec->codec_id = CODEC_ID_AAC; } } } @@ -316,15 +433,7 @@ static int mov_read_mdat(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { if(atom.size == 0) /* wrong one (MP4) */ return 0; - c->mdat_list = av_realloc(c->mdat_list, (c->mdat_count + 1) * sizeof(*c->mdat_list)); - c->mdat_list[c->mdat_count].offset = atom.offset; - c->mdat_list[c->mdat_count].size = atom.size; - c->mdat_count++; c->found_mdat=1; - c->mdat_offset = atom.offset; - if(c->found_moov) - return 1; /* found both, just go */ - url_fskip(pb, atom.size); return 0; /* now go for moov */ } @@ -348,11 +457,15 @@ static int mov_read_moov(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) /* we parsed the 'moov' atom, we can terminate the parsing as soon as we find the 'mdat' */ /* so we don't parse the whole file if over a network */ c->found_moov=1; - if(c->found_mdat) - return 1; /* found both, just go */ return 0; /* now go for mdat */ } +static int mov_read_moof(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) +{ + c->fragment.moof_offset = url_ftell(pb) - 8; + dprintf(c->fc, "moof offset %llx\n", c->fragment.moof_offset); + return mov_read_default(c, pb, atom); +} static int mov_read_mdhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { @@ -362,11 +475,9 @@ static int mov_read_mdhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) int lang; if (version > 1) - return 1; /* unsupported */ - - get_byte(pb); get_byte(pb); - get_byte(pb); /* flags */ + return -1; /* unsupported */ + get_be24(pb); /* flags */ if (version == 1) { get_be64(pb); get_be64(pb); @@ -388,7 +499,7 @@ static int mov_read_mdhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) static int mov_read_mvhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { int version = get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ if (version == 1) { get_be64(pb); @@ -431,16 +542,13 @@ static int mov_read_smi(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) // currently SVQ3 decoder expect full STSD header - so let's fake it // this should be fixed and just SMI header should be passed av_free(st->codec->extradata); + st->codec->extradata = av_mallocz(atom.size + 0x5a + FF_INPUT_BUFFER_PADDING_SIZE); + if (!st->codec->extradata) + return AVERROR(ENOMEM); st->codec->extradata_size = 0x5a + atom.size; - st->codec->extradata = av_mallocz(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE); - - if (st->codec->extradata) { - memcpy(st->codec->extradata, "SVQ3", 4); // fake - get_buffer(pb, st->codec->extradata + 0x5a, atom.size); - dprintf(c->fc, "Reading SMI %"PRId64" %s\n", atom.size, st->codec->extradata + 0x5a); - } else - url_fskip(pb, atom.size); - + memcpy(st->codec->extradata, "SVQ3", 4); // fake + get_buffer(pb, st->codec->extradata + 0x5a, atom.size); + dprintf(c->fc, "Reading SMI %"PRId64" %s\n", atom.size, st->codec->extradata + 0x5a); return 0; } @@ -494,13 +602,11 @@ static int mov_read_wave(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) if (st->codec->codec_id == CODEC_ID_QDM2) { // pass all frma atom to codec, needed at least for QDM2 av_free(st->codec->extradata); + st->codec->extradata = av_mallocz(atom.size + FF_INPUT_BUFFER_PADDING_SIZE); + if (!st->codec->extradata) + return AVERROR(ENOMEM); st->codec->extradata_size = atom.size; - st->codec->extradata = av_mallocz(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE); - - if (st->codec->extradata) { - get_buffer(pb, st->codec->extradata, atom.size); - } else - url_fskip(pb, atom.size); + get_buffer(pb, st->codec->extradata, atom.size); } else if (atom.size > 8) { /* to read frma, esds atoms */ if (mov_read_default(c, pb, atom) < 0) return -1; @@ -509,7 +615,11 @@ static int mov_read_wave(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) return 0; } -static int mov_read_avcC(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) +/** + * This function reads atom content and puts data in extradata without tag + * nor size unlike mov_read_extradata. + */ +static int mov_read_glbl(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { AVStream *st = c->fc->streams[c->fc->nb_streams-1]; @@ -517,15 +627,11 @@ static int mov_read_avcC(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) return -1; av_free(st->codec->extradata); - + st->codec->extradata = av_mallocz(atom.size + FF_INPUT_BUFFER_PADDING_SIZE); + if (!st->codec->extradata) + return AVERROR(ENOMEM); st->codec->extradata_size = atom.size; - st->codec->extradata = av_mallocz(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE); - - if (st->codec->extradata) { - get_buffer(pb, st->codec->extradata, atom.size); - } else - url_fskip(pb, atom.size); - + get_buffer(pb, st->codec->extradata, atom.size); return 0; } @@ -536,7 +642,7 @@ static int mov_read_stco(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) unsigned int i, entries; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ entries = get_be32(pb); @@ -547,15 +653,13 @@ static int mov_read_stco(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) sc->chunk_offsets = av_malloc(entries * sizeof(int64_t)); if (!sc->chunk_offsets) return -1; - if (atom.type == MKTAG('s', 't', 'c', 'o')) { - for(i=0; ichunk_offsets[i] = get_be32(pb); - } - } else if (atom.type == MKTAG('c', 'o', '6', '4')) { - for(i=0; ichunk_offsets[i] = get_be64(pb); - } - } else + else return -1; return 0; @@ -565,66 +669,71 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { AVStream *st = c->fc->streams[c->fc->nb_streams-1]; MOVStreamContext *sc = st->priv_data; - int entries, frames_per_sample; - uint32_t format; - uint8_t codec_name[32]; - - /* for palette traversal */ - unsigned int color_depth; - unsigned int color_start; - unsigned int color_count; - unsigned int color_end; - int color_index; - int color_dec; - int color_greyscale; - const uint8_t *color_table; - int j; - unsigned char r, g, b; + int j, entries, pseudo_stream_id; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ entries = get_be32(pb); - while(entries--) { //Parsing Sample description table + for(pseudo_stream_id=0; pseudo_stream_idcodec->codec_tag) { - /* multiple fourcc, just skip for now */ + dref_id = get_be16(pb); + + if (st->codec->codec_tag && + st->codec->codec_tag != format && + (c->fc->video_codec_id ? codec_get_id(codec_movvideo_tags, format) != c->fc->video_codec_id + : st->codec->codec_tag != MKTAG('j','p','e','g')) + ){ + /* Multiple fourcc, we skip JPEG. This is not correct, we should + * export it as a separate AVStream but this needs a few changes + * in the MOV demuxer, patch welcome. */ + av_log(c->fc, AV_LOG_WARNING, "multiple fourcc not supported\n"); url_fskip(pb, size - (url_ftell(pb) - start_pos)); continue; } + sc->pseudo_stream_id = st->codec->codec_tag ? -1 : pseudo_stream_id; + sc->dref_id= dref_id; st->codec->codec_tag = format; id = codec_get_id(codec_movaudio_tags, format); - if (id<=0 && (format&0xFFFF) == 'm' + ('s'<<8)) + if (id<=0 && (format&0xFFFF) == 'm'+('s'<<8)) id = codec_get_id(codec_wav_tags, bswap_32(format)&0xFFFF); if (st->codec->codec_type != CODEC_TYPE_VIDEO && id > 0) { st->codec->codec_type = CODEC_TYPE_AUDIO; } else if (st->codec->codec_type != CODEC_TYPE_AUDIO && /* do not overwrite codec type */ - format && format != MKTAG('m', 'p', '4', 's')) { /* skip old asf mpeg4 tag */ + format && format != MKTAG('m','p','4','s')) { /* skip old asf mpeg4 tag */ id = codec_get_id(codec_movvideo_tags, format); if (id <= 0) id = codec_get_id(codec_bmp_tags, format); if (id > 0) st->codec->codec_type = CODEC_TYPE_VIDEO; + else if(st->codec->codec_type == CODEC_TYPE_DATA){ + id = codec_get_id(ff_codec_movsubtitle_tags, format); + if(id > 0) + st->codec->codec_type = CODEC_TYPE_SUBTITLE; + } } - dprintf(c->fc, "size=%d 4CC= %c%c%c%c codec_type=%d\n", - size, - (format >> 0) & 0xff, (format >> 8) & 0xff, (format >> 16) & 0xff, (format >> 24) & 0xff, - st->codec->codec_type); + dprintf(c->fc, "size=%d 4CC= %c%c%c%c codec_type=%d\n", size, + (format >> 0) & 0xff, (format >> 8) & 0xff, (format >> 16) & 0xff, + (format >> 24) & 0xff, st->codec->codec_type); if(st->codec->codec_type==CODEC_TYPE_VIDEO) { + uint8_t codec_name[32]; + unsigned int color_depth; + int color_greyscale; + st->codec->codec_id = id; get_be16(pb); /* version */ get_be16(pb); /* revision level */ @@ -638,11 +747,9 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) get_be32(pb); /* horiz resolution */ get_be32(pb); /* vert resolution */ get_be32(pb); /* data size, always 0 */ - frames_per_sample = get_be16(pb); /* frames per samples */ + get_be16(pb); /* frames per samples */ - dprintf(c->fc, "frames/samples = %d\n", frames_per_sample); - - get_buffer(pb, codec_name, 32); /* codec name, pascal string (FIXME: true for mp4?) */ + get_buffer(pb, codec_name, 32); /* codec name, pascal string */ if (codec_name[0] <= 31) { memcpy(st->codec->codec_name, &codec_name[1],codec_name[0]); st->codec->codec_name[codec_name[0]] = 0; @@ -650,7 +757,8 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) st->codec->bits_per_sample = get_be16(pb); /* depth */ st->codec->color_table_id = get_be16(pb); /* colortable id */ - + dprintf(c->fc, "depth %d, ctab id %d\n", + st->codec->bits_per_sample, st->codec->color_table_id); /* figure out the palette situation */ color_depth = st->codec->bits_per_sample & 0x1F; color_greyscale = st->codec->bits_per_sample & 0x20; @@ -658,8 +766,14 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) /* if the depth is 2, 4, or 8 bpp, file is palettized */ if ((color_depth == 2) || (color_depth == 4) || (color_depth == 8)) { + /* for palette traversal */ + unsigned int color_start, color_count, color_end; + unsigned char r, g, b; + if (color_greyscale) { + int color_index, color_dec; /* compute the greyscale palette */ + st->codec->bits_per_sample = color_depth; color_count = 1 << color_depth; color_index = 255; color_dec = 256 / (color_count - 1); @@ -671,7 +785,8 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) if (color_index < 0) color_index = 0; } - } else if (st->codec->color_table_id & 0x08) { + } else if (st->codec->color_table_id) { + const uint8_t *color_table; /* if flag bit 3 is set, use the default palette */ color_count = 1 << color_depth; if (color_depth == 2) @@ -727,10 +842,8 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) st->codec->channels = get_be16(pb); /* channel count */ dprintf(c->fc, "audio channels %d\n", st->codec->channels); st->codec->bits_per_sample = get_be16(pb); /* sample size */ - /* do we need to force to 16 for AMR ? */ - /* handle specific s8 codec */ - get_be16(pb); /* compression id = 0*/ + sc->audio_cid = get_be16(pb); get_be16(pb); /* packet size = 0 */ st->codec->sample_rate = ((get_be32(pb) >> 16)); @@ -748,6 +861,23 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) else if (st->codec->bits_per_sample == 24) st->codec->codec_id = CODEC_ID_PCM_S24BE; break; + /* set values for old format before stsd version 1 appeared */ + case CODEC_ID_MACE3: + sc->samples_per_frame = 6; + sc->bytes_per_frame = 2*st->codec->channels; + break; + case CODEC_ID_MACE6: + sc->samples_per_frame = 6; + sc->bytes_per_frame = 1*st->codec->channels; + break; + case CODEC_ID_ADPCM_IMA_QT: + sc->samples_per_frame = 64; + sc->bytes_per_frame = 34*st->codec->channels; + break; + case CODEC_ID_GSM: + sc->samples_per_frame = 160; + sc->bytes_per_frame = 33; + break; default: break; } @@ -777,6 +907,8 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) st->codec->bits_per_sample = bits_per_sample; sc->sample_size = (bits_per_sample >> 3) * st->codec->channels; } + } else if(st->codec->codec_type==CODEC_TYPE_SUBTITLE){ + st->codec->codec_id= id; } else { /* other codec type, just skip (rtp, mp4s, tmcd ...) */ url_fskip(pb, size - (url_ftell(pb) - start_pos)); @@ -790,33 +922,11 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) url_fskip(pb, a.size); } - if(st->codec->codec_type==CODEC_TYPE_AUDIO && st->codec->sample_rate==0 && sc->time_scale>1) { + if(st->codec->codec_type==CODEC_TYPE_AUDIO && st->codec->sample_rate==0 && sc->time_scale>1) st->codec->sample_rate= sc->time_scale; - } /* special codec parameters handling */ switch (st->codec->codec_id) { -#ifdef CONFIG_H261_DECODER - case CODEC_ID_H261: -#endif -#ifdef CONFIG_H263_DECODER - case CODEC_ID_H263: -#endif -#ifdef CONFIG_MPEG4_DECODER - case CODEC_ID_MPEG4: -#endif - st->codec->width= 0; /* let decoder init width/height */ - st->codec->height= 0; - break; -#ifdef CONFIG_LIBFAAD - case CODEC_ID_AAC: -#endif -#ifdef CONFIG_VORBIS_DECODER - case CODEC_ID_VORBIS: -#endif - case CODEC_ID_MP3ON4: - st->codec->sample_rate= 0; /* let decoder init parameters properly */ - break; #ifdef CONFIG_DV_DEMUXER case CODEC_ID_DVAUDIO: c->dv_fctx = av_alloc_format_context(); @@ -834,7 +944,9 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) st->codec->sample_rate= 16000; st->codec->channels= 1; /* really needed */ break; + case CODEC_ID_QCELP: case CODEC_ID_AMR_NB: + st->codec->frame_size= sc->samples_per_frame; st->codec->sample_rate= 8000; st->codec->channels= 1; /* really needed */ break; @@ -843,10 +955,15 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) st->codec->codec_type = CODEC_TYPE_AUDIO; /* force type after stsd for m1a hdlr */ st->need_parsing = AVSTREAM_PARSE_FULL; break; + case CODEC_ID_GSM: case CODEC_ID_ADPCM_MS: case CODEC_ID_ADPCM_IMA_WAV: st->codec->block_align = sc->bytes_per_frame; break; + case CODEC_ID_ALAC: + if (st->codec->extradata_size == 36) + st->codec->frame_size = AV_RB32((st->codec->extradata+12)); + break; default: break; } @@ -861,7 +978,7 @@ static int mov_read_stsc(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) unsigned int i, entries; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ entries = get_be32(pb); @@ -889,7 +1006,7 @@ static int mov_read_stss(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) unsigned int i, entries; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ entries = get_be32(pb); @@ -917,7 +1034,7 @@ static int mov_read_stsz(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) unsigned int i, entries, sample_size; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ sample_size = get_be32(pb); if (!sc->sample_size) /* do not overwrite value computed in stsd */ @@ -935,10 +1052,8 @@ static int mov_read_stsz(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) sc->sample_sizes = av_malloc(entries * sizeof(int)); if (!sc->sample_sizes) return -1; - for(i=0; isample_sizes[i] = get_be32(pb); - dprintf(c->fc, "sample_sizes[]=%d\n", sc->sample_sizes[i]); - } return 0; } @@ -951,7 +1066,7 @@ static int mov_read_stts(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) int64_t total_sample_count=0; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ entries = get_be32(pb); if(entries >= UINT_MAX / sizeof(MOV_stts_t)) return -1; @@ -994,7 +1109,7 @@ static int mov_read_ctts(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) unsigned int i, entries; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ + get_be24(pb); /* flags */ entries = get_be32(pb); if(entries >= UINT_MAX / sizeof(MOV_stts_t)) return -1; @@ -1023,25 +1138,192 @@ static int mov_read_ctts(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) return 0; } +static void mov_build_index(MOVContext *mov, AVStream *st) +{ + MOVStreamContext *sc = st->priv_data; + offset_t current_offset; + int64_t current_dts = 0; + unsigned int stts_index = 0; + unsigned int stsc_index = 0; + unsigned int stss_index = 0; + unsigned int i, j; + + /* 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 current_sample = 0; + unsigned int stts_sample = 0; + unsigned int keyframe, sample_size; + unsigned int distance = 0; + int key_off = sc->keyframes && sc->keyframes[0] == 1; + + 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) + stsc_index++; + for (j = 0; j < sc->sample_to_chunk[stsc_index].count; j++) { + if (current_sample >= sc->sample_count) { + av_log(mov->fc, AV_LOG_ERROR, "wrong sample count\n"); + goto out; + } + keyframe = !sc->keyframe_count || current_sample+key_off == sc->keyframes[stss_index]; + if (keyframe) { + distance = 0; + if (stss_index + 1 < sc->keyframe_count) + stss_index++; + } + 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) { + 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; + distance++; + stts_sample++; + current_sample++; + if (stts_index + 1 < sc->stts_count && stts_sample == sc->stts_data[stts_index].count) { + stts_sample = 0; + stts_index++; + } + } + } + } else { /* read whole chunk */ + unsigned int chunk_samples, chunk_size, chunk_duration; + 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) + stsc_index++; + chunk_samples = sc->sample_to_chunk[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)) { + if (sc->samples_per_frame < 160) + chunk_size = chunk_samples * sc->bytes_per_frame / sc->samples_per_frame; + else { + chunk_size = sc->bytes_per_frame; + frames = chunk_samples / sc->samples_per_frame; + chunk_samples = sc->samples_per_frame; + } + } else + chunk_size = chunk_samples * sc->sample_size; + for (j = 0; j < frames; j++) { + av_add_index_entry(st, current_offset, current_dts, chunk_size, 0, AVINDEX_KEYFRAME); + /* get chunk duration */ + chunk_duration = 0; + while (chunk_samples > 0) { + if (chunk_samples < sc->stts_data[stts_index].count) { + chunk_duration += sc->stts_data[stts_index].duration * chunk_samples; + sc->stts_data[stts_index].count -= chunk_samples; + break; + } else { + chunk_duration += sc->stts_data[stts_index].duration * chunk_samples; + chunk_samples -= sc->stts_data[stts_index].count; + if (stts_index + 1 < sc->stts_count) + stts_index++; + } + } + current_offset += sc->bytes_per_frame; + dprintf(mov->fc, "AVIndex stream %d, chunk %d, offset %"PRIx64", dts %"PRId64", " + "size %d, duration %d\n", st->index, i, current_offset, current_dts, + chunk_size, chunk_duration); + assert(chunk_duration % sc->time_rate == 0); + current_dts += chunk_duration / sc->time_rate; + } + } + } + out: + /* adjust sample count to avindex entries */ + sc->sample_count = st->nb_index_entries; +} + static int mov_read_trak(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { AVStream *st; MOVStreamContext *sc; + int ret; st = av_new_stream(c->fc, c->fc->nb_streams); - if (!st) return -2; + if (!st) return AVERROR(ENOMEM); sc = av_mallocz(sizeof(MOVStreamContext)); - if (!sc) { - av_free(st); - return -1; - } + if (!sc) return AVERROR(ENOMEM); st->priv_data = sc; st->codec->codec_type = CODEC_TYPE_DATA; st->start_time = 0; /* XXX: check */ - c->streams[c->fc->nb_streams-1] = sc; - return mov_read_default(c, pb, atom); + 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))){ + 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; + 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); + + 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) { + if (url_fopen(&sc->pb, sc->drefs[sc->dref_id-1].path, URL_RDONLY) < 0) + av_log(c->fc, AV_LOG_ERROR, "stream %d, error opening file %s: %s\n", + st->index, sc->drefs[sc->dref_id-1].path, strerror(errno)); + } else + sc->pb = c->fc->pb; + + switch (st->codec->codec_id) { +#ifdef CONFIG_H261_DECODER + case CODEC_ID_H261: +#endif +#ifdef CONFIG_H263_DECODER + case CODEC_ID_H263: +#endif +#ifdef CONFIG_MPEG4_DECODER + case CODEC_ID_MPEG4: +#endif + st->codec->width= 0; /* let decoder init width/height */ + st->codec->height= 0; + break; +#ifdef CONFIG_VORBIS_DECODER + case CODEC_ID_VORBIS: +#endif + st->codec->sample_rate= 0; /* let decoder init parameters properly */ + break; + } + + /* Do not need those anymore. */ + av_freep(&sc->chunk_offsets); + av_freep(&sc->sample_to_chunk); + av_freep(&sc->sample_sizes); + av_freep(&sc->keyframes); + av_freep(&sc->stts_data); + + return 0; } static void mov_parse_udta_string(ByteIOContext *pb, char *str, int size) @@ -1092,8 +1374,7 @@ static int mov_read_tkhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) AVStream *st = c->fc->streams[c->fc->nb_streams-1]; int version = get_byte(pb); - get_byte(pb); get_byte(pb); - get_byte(pb); /* flags */ + get_be24(pb); /* flags */ /* MOV_TRACK_ENABLED 0x0001 MOV_TRACK_IN_MOVIE 0x0002 @@ -1111,7 +1392,8 @@ static int mov_read_tkhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) st->id = (int)get_be32(pb); /* track id (NOT 0 !)*/ get_be32(pb); /* reserved */ st->start_time = 0; /* check */ - (version == 1) ? get_be64(pb) : get_be32(pb); /* highlevel (considering edits) duration in movie timebase */ + /* highlevel (considering edits) duration in movie timebase */ + (version == 1) ? get_be64(pb) : get_be32(pb); get_be32(pb); /* reserved */ get_be32(pb); /* reserved */ @@ -1129,6 +1411,129 @@ static int mov_read_tkhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) return 0; } +static int mov_read_tfhd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) +{ + MOVFragment *frag = &c->fragment; + MOVTrackExt *trex = NULL; + int flags, track_id, i; + + get_byte(pb); /* version */ + flags = get_be24(pb); + + track_id = get_be32(pb); + if (!track_id || track_id > c->fc->nb_streams) + return -1; + frag->track_id = track_id; + for (i = 0; i < c->trex_count; i++) + if (c->trex_data[i].track_id == frag->track_id) { + trex = &c->trex_data[i]; + break; + } + if (!trex) { + av_log(c->fc, AV_LOG_ERROR, "could not find corresponding trex\n"); + return -1; + } + + if (flags & 0x01) frag->base_data_offset = get_be64(pb); + else frag->base_data_offset = frag->moof_offset; + if (flags & 0x02) frag->stsd_id = get_be32(pb); + else frag->stsd_id = trex->stsd_id; + + frag->duration = flags & 0x08 ? get_be32(pb) : trex->duration; + frag->size = flags & 0x10 ? get_be32(pb) : trex->size; + frag->flags = flags & 0x20 ? get_be32(pb) : trex->flags; + dprintf(c->fc, "frag flags 0x%x\n", frag->flags); + return 0; +} + +static int mov_read_trex(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) +{ + MOVTrackExt *trex; + + 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) + return AVERROR(ENOMEM); + trex = &c->trex_data[c->trex_count++]; + get_byte(pb); /* version */ + get_be24(pb); /* flags */ + trex->track_id = get_be32(pb); + trex->stsd_id = get_be32(pb); + trex->duration = get_be32(pb); + trex->size = get_be32(pb); + trex->flags = get_be32(pb); + return 0; +} + +static int mov_read_trun(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) +{ + MOVFragment *frag = &c->fragment; + AVStream *st; + MOVStreamContext *sc; + uint64_t offset; + int64_t dts; + int data_offset = 0; + unsigned entries, first_sample_flags = frag->flags; + int flags, distance, i; + + if (!frag->track_id || frag->track_id > c->fc->nb_streams) + 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; + get_byte(pb); /* version */ + flags = get_be24(pb); + entries = get_be32(pb); + dprintf(c->fc, "flags 0x%x entries %d\n", flags, entries); + if (flags & 0x001) data_offset = get_be32(pb); + if (flags & 0x004) first_sample_flags = get_be32(pb); + if (flags & 0x800) { + 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) + return AVERROR(ENOMEM); + } + dts = st->duration; + offset = frag->base_data_offset + data_offset; + distance = 0; + dprintf(c->fc, "first sample flags 0x%x\n", first_sample_flags); + for (i = 0; i < entries; i++) { + unsigned sample_size = frag->size; + int sample_flags = i ? frag->flags : first_sample_flags; + unsigned sample_duration = frag->duration; + int keyframe; + + if (flags & 0x100) sample_duration = get_be32(pb); + if (flags & 0x200) sample_size = get_be32(pb); + if (flags & 0x400) sample_flags = get_be32(pb); + if (flags & 0x800) { + sc->ctts_data[sc->ctts_count].count = 1; + sc->ctts_data[sc->ctts_count].duration = get_be32(pb); + sc->ctts_count++; + } + if ((keyframe = st->codec->codec_type == CODEC_TYPE_AUDIO || + (flags & 0x004 && !i && !sample_flags) || sample_flags & 0x2000000)) + distance = 0; + av_add_index_entry(st, offset, dts, sample_size, distance, + keyframe ? AVINDEX_KEYFRAME : 0); + dprintf(c->fc, "AVIndex stream %d, sample %d, offset %"PRIx64", dts %"PRId64", " + "size %d, distance %d, keyframe %d\n", st->index, sc->sample_count+i, + offset, dts, sample_size, distance, keyframe); + distance++; + assert(sample_duration % sc->time_rate == 0); + dts += sample_duration / sc->time_rate; + offset += sample_size; + } + frag->moof_offset = offset; + sc->sample_count = st->nb_index_entries; + st->duration = dts; + return 0; +} + /* this atom should be null (from specs), but some buggy files put the 'moov' atom inside it... */ /* like the files created with Adobe Premiere 5.0, for samples see */ /* http://graphics.tudelft.nl/~wouter/publications/soundtests/ */ @@ -1145,7 +1550,7 @@ static int mov_read_wide(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) atom.type = get_le32(pb); atom.offset += 8; atom.size -= 8; - if (atom.type != MKTAG('m', 'd', 'a', 't')) { + if (atom.type != MKTAG('m','d','a','t')) { url_fskip(pb, atom.size); return 0; } @@ -1160,17 +1565,17 @@ static int mov_read_cmov(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) uint8_t *cmov_data; uint8_t *moov_data; /* uncompressed data */ long cmov_len, moov_len; - int ret; + int ret = -1; get_be32(pb); /* dcom atom */ - if (get_le32(pb) != MKTAG( 'd', 'c', 'o', 'm' )) + if (get_le32(pb) != MKTAG('d','c','o','m')) return -1; - if (get_le32(pb) != MKTAG( 'z', 'l', 'i', 'b' )) { + if (get_le32(pb) != MKTAG('z','l','i','b')) { av_log(NULL, AV_LOG_ERROR, "unknown compression for cmov atom !"); return -1; } get_be32(pb); /* cmvd atom */ - if (get_le32(pb) != MKTAG( 'c', 'm', 'v', 'd' )) + if (get_le32(pb) != MKTAG('c','m','v','d')) return -1; moov_len = get_be32(pb); /* uncompressed size */ cmov_len = atom.size - 6 * 4; @@ -1185,16 +1590,17 @@ static int mov_read_cmov(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) } get_buffer(pb, cmov_data, cmov_len); if(uncompress (moov_data, (uLongf *) &moov_len, (const Bytef *)cmov_data, cmov_len) != Z_OK) - return -1; + goto free_and_return; if(init_put_byte(&ctx, moov_data, moov_len, 0, NULL, NULL, NULL, NULL) != 0) - return -1; - atom.type = MKTAG( 'm', 'o', 'o', 'v' ); + goto free_and_return; + atom.type = MKTAG('m','o','o','v'); atom.offset = 0; atom.size = moov_len; #ifdef DEBUG // { int fd = open("/tmp/uncompheader.mov", O_WRONLY | O_CREAT); write(fd, moov_data, moov_len); close(fd); } #endif ret = mov_read_default(c, &ctx, atom); +free_and_return: av_free(moov_data); av_free(cmov_data); return ret; @@ -1207,59 +1613,71 @@ static int mov_read_cmov(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) /* edit list atom */ static int mov_read_elst(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom) { + MOVStreamContext *sc = c->fc->streams[c->fc->nb_streams-1]->priv_data; int i, edit_count; get_byte(pb); /* version */ - get_byte(pb); get_byte(pb); get_byte(pb); /* flags */ - edit_count= c->streams[c->fc->nb_streams-1]->edit_count = get_be32(pb); /* entries */ + get_be24(pb); /* flags */ + edit_count= sc->edit_count = get_be32(pb); /* entries */ for(i=0; ifc, AV_LOG_WARNING, "edit list not starting at 0, " + "a/v desync might occur, patch welcome\n"); } - dprintf(c->fc, "track[%i].edit_count = %i\n", c->fc->nb_streams-1, c->streams[c->fc->nb_streams-1]->edit_count); + dprintf(c->fc, "track[%i].edit_count = %i\n", c->fc->nb_streams-1, sc->edit_count); return 0; } static const MOVParseTableEntry mov_default_parse_table[] = { -/* mp4 atoms */ -{ MKTAG( 'c', 'o', '6', '4' ), mov_read_stco }, -{ MKTAG( 'c', 't', 't', 's' ), mov_read_ctts }, /* composition time to sample */ -{ MKTAG( 'e', 'd', 't', 's' ), mov_read_default }, -{ MKTAG( 'e', 'l', 's', 't' ), mov_read_elst }, -{ MKTAG( 'e', 'n', 'd', 'a' ), mov_read_enda }, -{ MKTAG( 'f', 'i', 'e', 'l' ), mov_read_extradata }, -{ MKTAG( 'f', 't', 'y', 'p' ), mov_read_ftyp }, -{ MKTAG( 'h', 'd', 'l', 'r' ), mov_read_hdlr }, -{ MKTAG( 'j', 'p', '2', 'h' ), mov_read_extradata }, -{ MKTAG( 'm', 'd', 'a', 't' ), mov_read_mdat }, -{ MKTAG( 'm', 'd', 'h', 'd' ), mov_read_mdhd }, -{ MKTAG( 'm', 'd', 'i', 'a' ), mov_read_default }, -{ MKTAG( 'm', 'i', 'n', 'f' ), mov_read_default }, -{ MKTAG( 'm', 'o', 'o', 'v' ), mov_read_moov }, -{ MKTAG( 'm', 'v', 'h', 'd' ), mov_read_mvhd }, -{ MKTAG( 'S', 'M', 'I', ' ' ), mov_read_smi }, /* Sorenson extension ??? */ -{ MKTAG( 'a', 'l', 'a', 'c' ), mov_read_extradata }, /* alac specific atom */ -{ MKTAG( 'a', 'v', 'c', 'C' ), mov_read_avcC }, -{ MKTAG( 's', 't', 'b', 'l' ), mov_read_default }, -{ MKTAG( 's', 't', 'c', 'o' ), mov_read_stco }, -{ MKTAG( 's', 't', 's', 'c' ), mov_read_stsc }, -{ MKTAG( 's', 't', 's', 'd' ), mov_read_stsd }, /* sample description */ -{ 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( 't', 'k', 'h', 'd' ), mov_read_tkhd }, /* track header */ -{ MKTAG( 't', 'r', 'a', 'k' ), mov_read_trak }, -{ MKTAG( 'u', 'd', 't', 'a' ), mov_read_udta }, -{ 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 }, -{ 0L, NULL } +{ MKTAG('c','o','6','4'), mov_read_stco }, +{ 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('e','d','t','s'), mov_read_default }, +{ MKTAG('e','l','s','t'), mov_read_elst }, +{ MKTAG('e','n','d','a'), mov_read_enda }, +{ MKTAG('f','i','e','l'), mov_read_extradata }, +{ MKTAG('f','t','y','p'), mov_read_ftyp }, +{ MKTAG('g','l','b','l'), mov_read_glbl }, +{ MKTAG('h','d','l','r'), mov_read_hdlr }, +{ MKTAG('j','p','2','h'), mov_read_extradata }, +{ MKTAG('m','d','a','t'), mov_read_mdat }, +{ MKTAG('m','d','h','d'), mov_read_mdhd }, +{ MKTAG('m','d','i','a'), mov_read_default }, +{ MKTAG('m','i','n','f'), mov_read_default }, +{ MKTAG('m','o','o','f'), mov_read_moof }, +{ MKTAG('m','o','o','v'), mov_read_moov }, +{ MKTAG('m','v','e','x'), mov_read_default }, +{ MKTAG('m','v','h','d'), mov_read_mvhd }, +{ MKTAG('S','M','I',' '), mov_read_smi }, /* Sorenson extension ??? */ +{ MKTAG('a','l','a','c'), mov_read_extradata }, /* alac specific atom */ +{ MKTAG('a','v','c','C'), mov_read_glbl }, +{ MKTAG('s','t','b','l'), mov_read_default }, +{ MKTAG('s','t','c','o'), mov_read_stco }, +{ MKTAG('s','t','s','c'), mov_read_stsc }, +{ MKTAG('s','t','s','d'), mov_read_stsd }, /* sample description */ +{ 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('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','u','n'), mov_read_trun }, +{ MKTAG('u','d','t','a'), mov_read_udta }, +{ 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 }, +{ 0, NULL } }; -/* XXX: is it sufficient ? */ static int mov_probe(AVProbeData *p) { unsigned int offset; @@ -1275,22 +1693,24 @@ static int mov_probe(AVProbeData *p) tag = AV_RL32(p->buf + offset + 4); switch(tag) { /* check for obvious tags */ - case MKTAG( 'j', 'P', ' ', ' ' ): /* jpeg 2000 signature */ - case MKTAG( 'm', 'o', 'o', 'v' ): - case MKTAG( 'm', 'd', 'a', 't' ): - case MKTAG( 'p', 'n', 'o', 't' ): /* detect movs with preview pics like ew.mov and april.mov */ - case MKTAG( 'u', 'd', 't', 'a' ): /* Packet Video PVAuthor adds this and a lot of more junk */ + case MKTAG('j','P',' ',' '): /* jpeg 2000 signature */ + case MKTAG('m','o','o','v'): + case MKTAG('m','d','a','t'): + case MKTAG('p','n','o','t'): /* detect movs with preview pics like ew.mov and april.mov */ + case MKTAG('u','d','t','a'): /* Packet Video PVAuthor adds this and a lot of more junk */ return AVPROBE_SCORE_MAX; /* those are more common words, so rate then a bit less */ - case MKTAG( 'e', 'd', 'i', 'w' ): /* xdcam files have reverted first tags */ - case MKTAG( 'w', 'i', 'd', 'e' ): - case MKTAG( 'f', 'r', 'e', 'e' ): - case MKTAG( 'j', 'u', 'n', 'k' ): - case MKTAG( 'p', 'i', 'c', 't' ): + case MKTAG('e','d','i','w'): /* xdcam files have reverted first tags */ + case MKTAG('w','i','d','e'): + case MKTAG('f','r','e','e'): + case MKTAG('j','u','n','k'): + case MKTAG('p','i','c','t'): return AVPROBE_SCORE_MAX - 5; - case MKTAG( 'f', 't', 'y', 'p' ): - case MKTAG( 's', 'k', 'i', 'p' ): - case MKTAG( 'u', 'u', 'i', 'd' ): + case MKTAG(0x82,0x82,0x7f,0x7d ): + case MKTAG('f','t','y','p'): + case MKTAG('s','k','i','p'): + case MKTAG('u','u','i','d'): + case MKTAG('p','r','f','l'): offset = AV_RB32(p->buf+offset) + offset; /* if we only find those cause probedata is too small at least rate them */ score = AVPROBE_SCORE_MAX - 50; @@ -1303,179 +1723,31 @@ static int mov_probe(AVProbeData *p) return score; } -static void mov_build_index(MOVContext *mov, AVStream *st) -{ - MOVStreamContext *sc = st->priv_data; - offset_t current_offset; - int64_t current_dts = 0; - unsigned int stts_index = 0; - unsigned int stsc_index = 0; - unsigned int stss_index = 0; - unsigned int i, j, k; - - if (sc->sample_sizes || st->codec->codec_type == CODEC_TYPE_VIDEO || sc->dv_audio_container) { - unsigned int current_sample = 0; - unsigned int stts_sample = 0; - unsigned int keyframe, sample_size; - unsigned int distance = 0; - - 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) - stsc_index++; - for (j = 0; j < sc->sample_to_chunk[stsc_index].count; j++) { - if (current_sample >= sc->sample_count) { - av_log(mov->fc, AV_LOG_ERROR, "wrong sample count\n"); - goto out; - } - keyframe = !sc->keyframe_count || current_sample + 1 == sc->keyframes[stss_index]; - if (keyframe) { - distance = 0; - if (stss_index + 1 < sc->keyframe_count) - stss_index++; - } - sample_size = sc->sample_size > 0 ? sc->sample_size : sc->sample_sizes[current_sample]; - 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); - av_add_index_entry(st, current_offset, current_dts, sample_size, distance, keyframe ? AVINDEX_KEYFRAME : 0); - 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; - distance++; - stts_sample++; - current_sample++; - if (stts_index + 1 < sc->stts_count && stts_sample == sc->stts_data[stts_index].count) { - stts_sample = 0; - stts_index++; - } - } - } - } else { /* read whole chunk */ - unsigned int chunk_samples, chunk_size, chunk_duration; - 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) - stsc_index++; - chunk_samples = sc->sample_to_chunk[stsc_index].count; - /* get chunk size */ - if (sc->sample_size > 1 || st->codec->codec_id == CODEC_ID_PCM_U8 || st->codec->codec_id == CODEC_ID_PCM_S8) - chunk_size = chunk_samples * sc->sample_size; - else if (sc->samples_per_frame > 0 && (chunk_samples * sc->bytes_per_frame % sc->samples_per_frame == 0)) - chunk_size = chunk_samples * sc->bytes_per_frame / sc->samples_per_frame; - else { /* workaround to find nearest next chunk offset */ - chunk_size = INT_MAX; - for (j = 0; j < mov->total_streams; j++) { - MOVStreamContext *msc = mov->streams[j]; - for (k = msc->next_chunk; k < msc->chunk_count; k++) { - if (msc->chunk_offsets[k] > current_offset && msc->chunk_offsets[k] - current_offset < chunk_size) { - chunk_size = msc->chunk_offsets[k] - current_offset; - msc->next_chunk = k; - break; - } - } - } - /* check for last chunk */ - if (chunk_size == INT_MAX) - for (j = 0; j < mov->mdat_count; j++) { - dprintf(mov->fc, "mdat %d, offset %"PRIx64", size %"PRId64", current offset %"PRIx64"\n", - j, mov->mdat_list[j].offset, mov->mdat_list[j].size, current_offset); - if (mov->mdat_list[j].offset <= current_offset && mov->mdat_list[j].offset + mov->mdat_list[j].size > current_offset) - chunk_size = mov->mdat_list[j].offset + mov->mdat_list[j].size - current_offset; - } - assert(chunk_size != INT_MAX); - for (j = 0; j < mov->total_streams; j++) { - mov->streams[j]->next_chunk = 0; - } - } - av_add_index_entry(st, current_offset, current_dts, chunk_size, 0, AVINDEX_KEYFRAME); - /* get chunk duration */ - chunk_duration = 0; - while (chunk_samples > 0) { - if (chunk_samples < sc->stts_data[stts_index].count) { - chunk_duration += sc->stts_data[stts_index].duration * chunk_samples; - sc->stts_data[stts_index].count -= chunk_samples; - break; - } else { - chunk_duration += sc->stts_data[stts_index].duration * chunk_samples; - chunk_samples -= sc->stts_data[stts_index].count; - if (stts_index + 1 < sc->stts_count) { - stts_index++; - } - } - } - dprintf(mov->fc, "AVIndex stream %d, chunk %d, offset %"PRIx64", dts %"PRId64", size %d, duration %d\n", - st->index, i, current_offset, current_dts, chunk_size, chunk_duration); - assert(chunk_duration % sc->time_rate == 0); - current_dts += chunk_duration / sc->time_rate; - } - } - out: - /* adjust sample count to avindex entries */ - sc->sample_count = st->nb_index_entries; -} - static int mov_read_header(AVFormatContext *s, AVFormatParameters *ap) { MOVContext *mov = s->priv_data; ByteIOContext *pb = s->pb; - int i, err; + int err; MOV_atom_t atom = { 0, 0, 0 }; mov->fc = s; - - if(!url_is_streamed(pb)) /* .mov and .mp4 aren't streamable anyway (only progressive download if moov is before mdat) */ + /* .mov and .mp4 aren't streamable anyway (only progressive download if moov is before mdat) */ + if(!url_is_streamed(pb)) atom.size = url_fsize(pb); else atom.size = INT64_MAX; /* check MOV header */ - err = mov_read_default(mov, pb, atom); - if (err<0 || (!mov->found_moov && !mov->found_mdat)) { - av_log(s, AV_LOG_ERROR, "mov: header not found !!! (err:%d, moov:%d, mdat:%d) pos:%"PRId64"\n", - err, mov->found_moov, mov->found_mdat, url_ftell(pb)); - return -1; + if ((err = mov_read_default(mov, pb, atom)) < 0) { + av_log(s, AV_LOG_ERROR, "error reading header: %d\n", err); + return err; } - dprintf(mov->fc, "on_parse_exit_offset=%d\n", (int) url_ftell(pb)); - - mov->total_streams = s->nb_streams; - - for(i=0; itotal_streams; i++) { - MOVStreamContext *sc = mov->streams[i]; - AVStream *st = s->streams[i]; - /* sanity checks */ - if(!sc->stts_count || !sc->chunk_count || !sc->sample_to_chunk_sz || - (!sc->sample_size && !sc->sample_count)){ - av_log(s, AV_LOG_ERROR, "missing mandatory atoms, broken header\n"); - sc->sample_count = 0; //ignore track - continue; - } - if(!sc->time_rate) - sc->time_rate=1; - if(!sc->time_scale) - sc->time_scale= mov->time_scale; - av_set_pts_info(st, 64, sc->time_rate, sc->time_scale); - - if (st->codec->codec_type == CODEC_TYPE_AUDIO && sc->stts_count == 1) - st->codec->frame_size = sc->stts_data[0].duration; - - if(st->duration != AV_NOPTS_VALUE){ - assert(st->duration % sc->time_rate == 0); - st->duration /= sc->time_rate; - } - sc->ffindex = i; - mov_build_index(mov, st); + if (!mov->found_moov) { + av_log(s, AV_LOG_ERROR, "moov atom not found\n"); + return -1; } + dprintf(mov->fc, "on_parse_exit_offset=%lld\n", url_ftell(pb)); - for(i=0; itotal_streams; i++) { - /* Do not need those anymore. */ - av_freep(&mov->streams[i]->chunk_offsets); - av_freep(&mov->streams[i]->sample_to_chunk); - av_freep(&mov->streams[i]->sample_sizes); - av_freep(&mov->streams[i]->keyframes); - av_freep(&mov->streams[i]->stts_data); - } - av_freep(&mov->mdat_list); return 0; } @@ -1486,44 +1758,50 @@ static int mov_read_packet(AVFormatContext *s, AVPacket *pkt) AVIndexEntry *sample = 0; int64_t best_dts = INT64_MAX; int i; - - for (i = 0; i < mov->total_streams; i++) { - MOVStreamContext *msc = mov->streams[i]; - if (s->streams[i]->discard != AVDISCARD_ALL && msc->current_sample < msc->sample_count) { - AVIndexEntry *current_sample = &s->streams[i]->index_entries[msc->current_sample]; - int64_t dts = av_rescale(current_sample->timestamp * (int64_t)msc->time_rate, AV_TIME_BASE, msc->time_scale); + 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) { + 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); dprintf(s, "stream %d, sample %d, dts %"PRId64"\n", i, msc->current_sample, dts); if (!sample || (url_is_streamed(s->pb) && current_sample->pos < sample->pos) || (!url_is_streamed(s->pb) && + ((msc->pb != s->pb && dts < best_dts) || (msc->pb == s->pb && ((FFABS(best_dts - dts) <= AV_TIME_BASE && current_sample->pos < sample->pos) || - (FFABS(best_dts - dts) > AV_TIME_BASE && dts < best_dts)))) { + (FFABS(best_dts - dts) > AV_TIME_BASE && dts < best_dts)))))) { sample = current_sample; best_dts = dts; sc = msc; } } } - if (!sample) - return -1; + if (!sample) { + mov->found_mdat = 0; + if (!url_is_streamed(s->pb) || + mov_read_default(mov, s->pb, (MOV_atom_t){ 0, 0, INT64_MAX }) < 0 || + url_feof(s->pb)) + return -1; + dprintf(s, "read fragments, offset 0x%llx\n", url_ftell(s->pb)); + goto retry; + } /* must be done just before reading, to avoid infinite loop on sample */ sc->current_sample++; - if (url_fseek(s->pb, sample->pos, SEEK_SET) != sample->pos) { - av_log(mov->fc, AV_LOG_ERROR, "stream %d, offset 0x%"PRIx64": partial file\n", sc->ffindex, sample->pos); + if (url_fseek(sc->pb, sample->pos, SEEK_SET) != sample->pos) { + av_log(mov->fc, AV_LOG_ERROR, "stream %d, offset 0x%"PRIx64": partial file\n", + sc->ffindex, sample->pos); return -1; } + av_get_packet(sc->pb, pkt, sample->size); #ifdef CONFIG_DV_DEMUXER - if (sc->dv_audio_container) { - dv_get_packet(mov->dv_demux, pkt); - dprintf(s, "dv audio pkt size %d\n", pkt->size); - } else { -#endif - av_get_packet(s->pb, pkt, sample->size); -#ifdef CONFIG_DV_DEMUXER - if (mov->dv_demux) { - void *pkt_destruct_func = pkt->destruct; - dv_produce_packet(mov->dv_demux, pkt, pkt->data, pkt->size); - pkt->destruct = pkt_destruct_func; - } + if (mov->dv_demux && sc->dv_audio_container) { + dv_produce_packet(mov->dv_demux, pkt, pkt->data, pkt->size); + av_free(pkt->data); + pkt->size = 0; + if (dv_get_packet(mov->dv_demux, pkt) < 0) + return -1; } #endif pkt->stream_index = sc->ffindex; @@ -1533,16 +1811,22 @@ static int mov_read_packet(AVFormatContext *s, AVPacket *pkt) pkt->pts = pkt->dts + sc->ctts_data[sc->sample_to_ctime_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) { + 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; } } else { + AVStream *st = s->streams[sc->ffindex]; + int64_t next_dts = (sc->current_sample < sc->sample_count) ? + st->index_entries[sc->current_sample].timestamp : st->duration; + pkt->duration = next_dts - pkt->dts; pkt->pts = pkt->dts; } pkt->flags |= sample->flags & AVINDEX_KEYFRAME ? PKT_FLAG_KEY : 0; pkt->pos = sample->pos; - dprintf(s, "stream %d, pts %"PRId64", dts %"PRId64", pos 0x%"PRIx64", duration %d\n", pkt->stream_index, pkt->pts, pkt->dts, pkt->pos, pkt->duration); + dprintf(s, "stream %d, pts %"PRId64", dts %"PRId64", pos 0x%"PRIx64", duration %d\n", + pkt->stream_index, pkt->pts, pkt->dts, pkt->pos, pkt->duration); return 0; } @@ -1605,11 +1889,16 @@ static int mov_read_seek(AVFormatContext *s, int stream_index, int64_t sample_ti static int mov_read_close(AVFormatContext *s) { - int i; + int i, j; MOVContext *mov = s->priv_data; - for(i=0; itotal_streams; i++) { - av_freep(&mov->streams[i]->ctts_data); - av_freep(&mov->streams[i]); + for(i=0; inb_streams; i++) { + MOVStreamContext *sc = s->streams[i]->priv_data; + av_freep(&sc->ctts_data); + for (j=0; jdrefs_count; j++) + av_freep(&sc->drefs[j].path); + av_freep(&sc->drefs); + if (sc->pb && sc->pb != s->pb) + url_fclose(sc->pb); } if(mov->dv_demux){ for(i=0; idv_fctx->nb_streams; i++){ @@ -1619,12 +1908,13 @@ static int mov_read_close(AVFormatContext *s) av_freep(&mov->dv_fctx); av_freep(&mov->dv_demux); } + av_freep(&mov->trex_data); return 0; } AVInputFormat mov_demuxer = { "mov,mp4,m4a,3gp,3g2,mj2", - "QuickTime/MPEG4/Motion JPEG 2000 format", + NULL_IF_CONFIG_SMALL("QuickTime/MPEG-4/Motion JPEG 2000 format"), sizeof(MOVContext), mov_probe, mov_read_header,