diff options
Diffstat (limited to 'libavformat/matroskaenc.c')
| -rw-r--r-- | libavformat/matroskaenc.c | 846 |
1 files changed, 562 insertions, 284 deletions
diff --git a/libavformat/matroskaenc.c b/libavformat/matroskaenc.c index cab43f499e..80f7d7f9c3 100644 --- a/libavformat/matroskaenc.c +++ b/libavformat/matroskaenc.c @@ -2,20 +2,20 @@ * Matroska muxer * Copyright (c) 2007 David Conrad * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,12 +24,14 @@ #include "avc.h" #include "hevc.h" #include "avformat.h" +#include "avio_internal.h" #include "avlanguage.h" #include "flacenc.h" #include "internal.h" #include "isom.h" #include "matroska.h" #include "riff.h" +#include "subtitles.h" #include "vorbiscomment.h" #include "wv.h" @@ -43,10 +45,12 @@ #include "libavutil/opt.h" #include "libavutil/random_seed.h" #include "libavutil/samplefmt.h" +#include "libavutil/sha.h" #include "libavutil/stereo3d.h" #include "libavcodec/xiph.h" #include "libavcodec/mpeg4audio.h" +#include "libavcodec/internal.h" typedef struct ebml_master { int64_t pos; ///< absolute offset in the file where the master's elements start @@ -69,8 +73,11 @@ typedef struct mkv_seekhead { typedef struct mkv_cuepoint { uint64_t pts; + int stream_idx; int tracknum; int64_t cluster_pos; ///< file offset of the cluster containing the block + int64_t relative_pos; ///< relative offset from the position of the cluster containing the block + int64_t duration; ///< duration of the block according to time base } mkv_cuepoint; typedef struct mkv_cues { @@ -81,6 +88,7 @@ typedef struct mkv_cues { typedef struct mkv_track { int write_dts; + int has_cue; int64_t ts_offset; } mkv_track; @@ -110,7 +118,13 @@ typedef struct MatroskaMuxContext { int cluster_size_limit; int64_t cues_pos; int64_t cluster_time_limit; + int is_dash; + int dash_track_number; + + uint32_t chapter_id_offset; int wrote_chapters; + + int allow_raw_vfw; } MatroskaMuxContext; @@ -118,13 +132,16 @@ typedef struct MatroskaMuxContext { * offset, 4 bytes for target EBML ID */ #define MAX_SEEKENTRY_SIZE 21 -/** per-cuepoint-track - 3 1-byte EBML IDs, 3 1-byte EBML sizes, 2 +/** per-cuepoint-track - 5 1-byte EBML IDs, 5 1-byte EBML sizes, 4 * 8-byte uint max */ -#define MAX_CUETRACKPOS_SIZE 22 +#define MAX_CUETRACKPOS_SIZE 42 /** per-cuepoint - 2 1-byte EBML IDs, 2 1-byte EBML sizes, 8-byte uint max */ #define MAX_CUEPOINT_SIZE(num_tracks) 12 + MAX_CUETRACKPOS_SIZE * num_tracks +/** Seek preroll value for opus */ +#define OPUS_SEEK_PREROLL 80000000 + static int ebml_id_size(unsigned int id) { return (av_log2(id + 1) - 1) / 7 + 1; @@ -134,7 +151,7 @@ static void put_ebml_id(AVIOContext *pb, unsigned int id) { int i = ebml_id_size(id); while (i--) - avio_w8(pb, id >> (i * 8)); + avio_w8(pb, (uint8_t)(id >> (i * 8))); } /** @@ -144,10 +161,9 @@ static void put_ebml_id(AVIOContext *pb, unsigned int id) */ static void put_ebml_size_unknown(AVIOContext *pb, int bytes) { - assert(bytes <= 8); + av_assert0(bytes <= 8); avio_w8(pb, 0x1ff >> bytes); - while (--bytes) - avio_w8(pb, 0xff); + ffio_fill(pb, 0xff, bytes - 1); } /** @@ -172,18 +188,18 @@ static void put_ebml_num(AVIOContext *pb, uint64_t num, int bytes) int i, needed_bytes = ebml_num_size(num); // sizes larger than this are currently undefined in EBML - assert(num < (1ULL << 56) - 1); + av_assert0(num < (1ULL << 56) - 1); if (bytes == 0) // don't care how many bytes are used, so use the min bytes = needed_bytes; // the bytes needed to write the given size would exceed the bytes // that we need to use, so write unknown size. This shouldn't happen. - assert(bytes >= needed_bytes); + av_assert0(bytes >= needed_bytes); num |= 1ULL << bytes * 7; for (i = bytes - 1; i >= 0; i--) - avio_w8(pb, num >> i * 8); + avio_w8(pb, (uint8_t)(num >> i * 8)); } static void put_ebml_uint(AVIOContext *pb, unsigned int elementid, uint64_t val) @@ -196,7 +212,20 @@ static void put_ebml_uint(AVIOContext *pb, unsigned int elementid, uint64_t val) put_ebml_id(pb, elementid); put_ebml_num(pb, bytes, 0); for (i = bytes - 1; i >= 0; i--) - avio_w8(pb, val >> i * 8); + avio_w8(pb, (uint8_t)(val >> i * 8)); +} + +static void put_ebml_sint(AVIOContext *pb, unsigned int elementid, int64_t val) +{ + int i, bytes = 1; + uint64_t tmp = 2*(val < 0 ? val^-1 : val); + + while (tmp>>=8) bytes++; + + put_ebml_id(pb, elementid); + put_ebml_num(pb, bytes, 0); + for (i = bytes - 1; i >= 0; i--) + avio_w8(pb, (uint8_t)(val >> i * 8)); } static void put_ebml_float(AVIOContext *pb, unsigned int elementid, double val) @@ -230,7 +259,7 @@ static void put_ebml_void(AVIOContext *pb, uint64_t size) { int64_t currentpos = avio_tell(pb); - assert(size >= 2); + av_assert0(size >= 2); put_ebml_id(pb, EBML_ID_VOID); // we need to subtract the length needed to store the size from the @@ -240,8 +269,7 @@ static void put_ebml_void(AVIOContext *pb, uint64_t size) put_ebml_num(pb, size - 1, 0); else put_ebml_num(pb, size - 9, 8); - while (avio_tell(pb) < currentpos + size) - avio_w8(pb, 0); + ffio_fill(pb, 0, currentpos + size - avio_tell(pb)); } static ebml_master start_ebml_master(AVIOContext *pb, unsigned int elementid, @@ -265,9 +293,7 @@ static void end_ebml_master(AVIOContext *pb, ebml_master master) static void put_xiph_size(AVIOContext *pb, int size) { - int i; - for (i = 0; i < size / 255; i++) - avio_w8(pb, 255); + ffio_fill(pb, 255, size / 255); avio_w8(pb, size % 255); } @@ -305,17 +331,16 @@ static mkv_seekhead *mkv_start_seekhead(AVIOContext *pb, int64_t segment_offset, static int mkv_add_seekhead_entry(mkv_seekhead *seekhead, unsigned int elementid, uint64_t filepos) { - int err; + mkv_seekhead_entry *entries = seekhead->entries; // don't store more elements than we reserved space for if (seekhead->max_entries > 0 && seekhead->max_entries <= seekhead->num_entries) return -1; - if ((err = av_reallocp_array(&seekhead->entries, seekhead->num_entries + 1, - sizeof(*seekhead->entries))) < 0) { - seekhead->num_entries = 0; - return err; - } + entries = av_realloc_array(entries, seekhead->num_entries + 1, sizeof(mkv_seekhead_entry)); + if (!entries) + return AVERROR(ENOMEM); + seekhead->entries = entries; seekhead->entries[seekhead->num_entries].elementid = elementid; seekhead->entries[seekhead->num_entries++].segmentpos = filepos - seekhead->segment_offset; @@ -370,7 +395,7 @@ static int64_t mkv_write_seekhead(AVIOContext *pb, mkv_seekhead *seekhead) currentpos = seekhead->filepos; } fail: - av_free(seekhead->entries); + av_freep(&seekhead->entries); av_free(seekhead); return currentpos; @@ -386,28 +411,32 @@ static mkv_cues *mkv_start_cues(int64_t segment_offset) return cues; } -static int mkv_add_cuepoint(mkv_cues *cues, int stream, int64_t ts, int64_t cluster_pos) +static int mkv_add_cuepoint(mkv_cues *cues, int stream, int tracknum, int64_t ts, + int64_t cluster_pos, int64_t relative_pos, int64_t duration) { - int err; + mkv_cuepoint *entries = cues->entries; if (ts < 0) return 0; - if ((err = av_reallocp_array(&cues->entries, cues->num_entries + 1, - sizeof(*cues->entries))) < 0) { - cues->num_entries = 0; - return err; - } + entries = av_realloc_array(entries, cues->num_entries + 1, sizeof(mkv_cuepoint)); + if (!entries) + return AVERROR(ENOMEM); + cues->entries = entries; cues->entries[cues->num_entries].pts = ts; - cues->entries[cues->num_entries].tracknum = stream + 1; - cues->entries[cues->num_entries++].cluster_pos = cluster_pos - cues->segment_offset; + cues->entries[cues->num_entries].stream_idx = stream; + cues->entries[cues->num_entries].tracknum = tracknum; + cues->entries[cues->num_entries].cluster_pos = cluster_pos - cues->segment_offset; + cues->entries[cues->num_entries].relative_pos = relative_pos; + cues->entries[cues->num_entries++].duration = duration; return 0; } -static int64_t mkv_write_cues(AVIOContext *pb, mkv_cues *cues, int num_tracks) +static int64_t mkv_write_cues(AVFormatContext *s, mkv_cues *cues, mkv_track *tracks, int num_tracks) { + AVIOContext *pb = s->pb; ebml_master cues_element; int64_t currentpos; int i, j; @@ -419,16 +448,39 @@ static int64_t mkv_write_cues(AVIOContext *pb, mkv_cues *cues, int num_tracks) ebml_master cuepoint, track_positions; mkv_cuepoint *entry = &cues->entries[i]; uint64_t pts = entry->pts; + int ctp_nb = 0; + + // Calculate the number of entries, so we know the element size + for (j = 0; j < num_tracks; j++) + tracks[j].has_cue = 0; + for (j = 0; j < cues->num_entries - i && entry[j].pts == pts; j++) { + int tracknum = entry[j].stream_idx; + av_assert0(tracknum>=0 && tracknum<num_tracks); + if (tracks[tracknum].has_cue && s->streams[tracknum]->codec->codec_type != AVMEDIA_TYPE_SUBTITLE) + continue; + tracks[tracknum].has_cue = 1; + ctp_nb ++; + } - cuepoint = start_ebml_master(pb, MATROSKA_ID_POINTENTRY, MAX_CUEPOINT_SIZE(num_tracks)); + cuepoint = start_ebml_master(pb, MATROSKA_ID_POINTENTRY, MAX_CUEPOINT_SIZE(ctp_nb)); put_ebml_uint(pb, MATROSKA_ID_CUETIME, pts); // put all the entries from different tracks that have the exact same // timestamp into the same CuePoint + for (j = 0; j < num_tracks; j++) + tracks[j].has_cue = 0; for (j = 0; j < cues->num_entries - i && entry[j].pts == pts; j++) { + int tracknum = entry[j].stream_idx; + av_assert0(tracknum>=0 && tracknum<num_tracks); + if (tracks[tracknum].has_cue && s->streams[tracknum]->codec->codec_type != AVMEDIA_TYPE_SUBTITLE) + continue; + tracks[tracknum].has_cue = 1; track_positions = start_ebml_master(pb, MATROSKA_ID_CUETRACKPOSITION, MAX_CUETRACKPOS_SIZE); - put_ebml_uint(pb, MATROSKA_ID_CUETRACK , entry[j].tracknum ); - put_ebml_uint(pb, MATROSKA_ID_CUECLUSTERPOSITION, entry[j].cluster_pos); + put_ebml_uint(pb, MATROSKA_ID_CUETRACK , entry[j].tracknum ); + put_ebml_uint(pb, MATROSKA_ID_CUECLUSTERPOSITION , entry[j].cluster_pos); + put_ebml_uint(pb, MATROSKA_ID_CUERELATIVEPOSITION, entry[j].relative_pos); + if (entry[j].duration != -1) + put_ebml_uint(pb, MATROSKA_ID_CUEDURATION , entry[j].duration); end_ebml_master(pb, track_positions); } i += j - 1; @@ -441,7 +493,7 @@ static int64_t mkv_write_cues(AVIOContext *pb, mkv_cues *cues, int num_tracks) static int put_xiph_codecpriv(AVFormatContext *s, AVIOContext *pb, AVCodecContext *codec) { - uint8_t *header_start[3]; + const uint8_t *header_start[3]; int header_len[3]; int first_header_size; int j; @@ -490,7 +542,7 @@ static int put_flac_codecpriv(AVFormatContext *s, if (write_comment) { const char *vendor = (s->flags & AVFMT_FLAG_BITEXACT) ? - "Libav" : LIBAVFORMAT_IDENT; + "Lavf" : LIBAVFORMAT_IDENT; AVDictionary *dict = NULL; uint8_t buf[32], *data, *p; int len; @@ -566,8 +618,11 @@ static int mkv_write_native_codecprivate(AVFormatContext *s, codec->extradata_size - 12); break; default: - if (codec->extradata_size) - avio_write(dyn_cp, codec->extradata, codec->extradata_size); + if (codec->codec_id == AV_CODEC_ID_PRORES && + ff_codec_get_id(ff_codec_movvideo_tags, codec->codec_tag) == AV_CODEC_ID_PRORES) { + avio_wl32(dyn_cp, codec->codec_tag); + } else if (codec->extradata_size && codec->codec_id != AV_CODEC_ID_TTA) + avio_write(dyn_cp, codec->extradata, codec->extradata_size); } return 0; @@ -592,30 +647,46 @@ static int mkv_write_codecprivate(AVFormatContext *s, AVIOContext *pb, if (!codec->codec_tag) codec->codec_tag = ff_codec_get_tag(ff_codec_movvideo_tags, codec->codec_id); - if (codec->extradata_size) + if (codec->extradata_size) { + if ( ff_codec_get_id(ff_codec_movvideo_tags, codec->codec_tag) == codec->codec_id + && ff_codec_get_id(ff_codec_movvideo_tags, AV_RL32(codec->extradata + 4)) != codec->codec_id + ) { + int i; + avio_wb32(dyn_cp, 0x5a + codec->extradata_size); + avio_wl32(dyn_cp, codec->codec_tag); + for(i = 0; i < 0x5a - 8; i++) + avio_w8(dyn_cp, 0); + } avio_write(dyn_cp, codec->extradata, codec->extradata_size); + } } else { + if (!ff_codec_get_tag(ff_codec_bmp_tags, codec->codec_id)) + av_log(s, AV_LOG_WARNING, "codec %s is not supported by this format\n", + avcodec_get_name(codec->codec_id)); + if (!codec->codec_tag) codec->codec_tag = ff_codec_get_tag(ff_codec_bmp_tags, codec->codec_id); - if (!codec->codec_tag) { - av_log(s, AV_LOG_ERROR, "No bmp codec ID found.\n"); - ret = -1; + if (!codec->codec_tag && codec->codec_id != AV_CODEC_ID_RAWVIDEO) { + av_log(s, AV_LOG_ERROR, "No bmp codec tag found for codec %s\n", + avcodec_get_name(codec->codec_id)); + ret = AVERROR(EINVAL); } - ff_put_bmp_header(dyn_cp, codec, ff_codec_bmp_tags, 0); + ff_put_bmp_header(dyn_cp, codec, ff_codec_bmp_tags, 0, 0); } } else if (codec->codec_type == AVMEDIA_TYPE_AUDIO) { unsigned int tag; tag = ff_codec_get_tag(ff_codec_wav_tags, codec->codec_id); if (!tag) { - av_log(s, AV_LOG_ERROR, "No wav codec ID found.\n"); - ret = -1; + av_log(s, AV_LOG_ERROR, "No wav codec tag found for codec %s\n", + avcodec_get_name(codec->codec_id)); + ret = AVERROR(EINVAL); } if (!codec->codec_tag) codec->codec_tag = tag; - ff_put_wav_header(dyn_cp, codec); + ff_put_wav_header(dyn_cp, codec, FF_PUT_WAV_HEADER_FORCE_WAVEFORMATEX); } codecpriv_size = avio_close_dyn_buf(dyn_cp, &codecpriv); @@ -626,19 +697,28 @@ static int mkv_write_codecprivate(AVFormatContext *s, AVIOContext *pb, return ret; } + static int mkv_write_stereo_mode(AVFormatContext *s, AVIOContext *pb, - AVStream *st, int mode) + AVStream *st, int mode, int *h_width, int *h_height) { int i; - int display_width, display_height; - int h_width = 1, h_height = 1; - AVCodecContext *codec = st->codec; + int ret = 0; AVDictionaryEntry *tag; MatroskaVideoStereoModeType format = MATROSKA_VIDEO_STEREOMODE_TYPE_NB; + *h_width = 1; + *h_height = 1; // convert metadata into proper side data and add it to the stream - if ((tag = av_dict_get(s->metadata, "stereo_mode", NULL, 0))) { + if ((tag = av_dict_get(st->metadata, "stereo_mode", NULL, 0)) || + (tag = av_dict_get( s->metadata, "stereo_mode", NULL, 0))) { int stereo_mode = atoi(tag->value); + + for (i=0; i<MATROSKA_VIDEO_STEREOMODE_TYPE_NB; i++) + if (!strcmp(tag->value, ff_matroska_video_stereo_mode[i])){ + stereo_mode = i; + break; + } + if (stereo_mode < MATROSKA_VIDEO_STEREOMODE_TYPE_NB && stereo_mode != 10 && stereo_mode != 12) { int ret = ff_mkv_stereo3d_conv(st, stereo_mode); @@ -661,13 +741,13 @@ static int mkv_write_stereo_mode(AVFormatContext *s, AVIOContext *pb, format = (stereo->flags & AV_STEREO3D_FLAG_INVERT) ? MATROSKA_VIDEO_STEREOMODE_TYPE_RIGHT_LEFT : MATROSKA_VIDEO_STEREOMODE_TYPE_LEFT_RIGHT; - h_width = 2; + *h_width = 2; break; case AV_STEREO3D_TOPBOTTOM: format = MATROSKA_VIDEO_STEREOMODE_TYPE_TOP_BOTTOM; if (stereo->flags & AV_STEREO3D_FLAG_INVERT) format--; - h_height = 2; + *h_height = 2; break; case AV_STEREO3D_CHECKERBOARD: format = MATROSKA_VIDEO_STEREOMODE_TYPE_CHECKERBOARD_LR; @@ -678,13 +758,13 @@ static int mkv_write_stereo_mode(AVFormatContext *s, AVIOContext *pb, format = MATROSKA_VIDEO_STEREOMODE_TYPE_ROW_INTERLEAVED_LR; if (stereo->flags & AV_STEREO3D_FLAG_INVERT) format--; - h_height = 2; + *h_height = 2; break; case AV_STEREO3D_COLUMNS: format = MATROSKA_VIDEO_STEREOMODE_TYPE_COL_INTERLEAVED_LR; if (stereo->flags & AV_STEREO3D_FLAG_INVERT) format--; - h_width = 2; + *h_width = 2; break; case AV_STEREO3D_FRAMESEQUENCE: format = MATROSKA_VIDEO_STEREOMODE_TYPE_BOTH_EYES_BLOCK_LR; @@ -692,40 +772,32 @@ static int mkv_write_stereo_mode(AVFormatContext *s, AVIOContext *pb, format++; break; } - break; } } + if (format == MATROSKA_VIDEO_STEREOMODE_TYPE_NB) + return ret; + // if webm, do not write unsupported modes - if (mode == MODE_WEBM && - (format > MATROSKA_VIDEO_STEREOMODE_TYPE_TOP_BOTTOM && - format != MATROSKA_VIDEO_STEREOMODE_TYPE_RIGHT_LEFT)) + if ((mode == MODE_WEBM && + format > MATROSKA_VIDEO_STEREOMODE_TYPE_TOP_BOTTOM && + format != MATROSKA_VIDEO_STEREOMODE_TYPE_RIGHT_LEFT) + || format >= MATROSKA_VIDEO_STEREOMODE_TYPE_NB) { + av_log(s, AV_LOG_ERROR, + "The specified stereo mode is not valid.\n"); format = MATROSKA_VIDEO_STEREOMODE_TYPE_NB; + return AVERROR(EINVAL); + } // write StereoMode if format is valid - if (format < MATROSKA_VIDEO_STEREOMODE_TYPE_NB) - put_ebml_uint(pb, MATROSKA_ID_VIDEOSTEREOMODE, format); + put_ebml_uint(pb, MATROSKA_ID_VIDEOSTEREOMODE, format); - // write DisplayWidth and DisplayHeight, they contain the size of - // a single source view and/or the display aspect ratio - display_width = codec->width / h_width; - display_height = codec->height / h_height; - if (st->sample_aspect_ratio.num) { - display_width *= av_q2d(st->sample_aspect_ratio); - put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYUNIT, 3); // DAR - } - if (st->sample_aspect_ratio.num || - format < MATROSKA_VIDEO_STEREOMODE_TYPE_NB) { - put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYWIDTH, display_width); - put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYHEIGHT, display_height); - } - - return 0; + return ret; } static int mkv_write_track(AVFormatContext *s, MatroskaMuxContext *mkv, - int i, AVIOContext *pb) + int i, AVIOContext *pb, int default_stream_exists) { AVStream *st = s->streams[i]; AVCodecContext *codec = st->codec; @@ -735,6 +807,8 @@ static int mkv_write_track(AVFormatContext *s, MatroskaMuxContext *mkv, int bit_depth = av_get_bits_per_sample(codec->codec_id); int sample_rate = codec->sample_rate; int output_sample_rate = 0; + int display_width_div = 1; + int display_height_div = 1; int j, ret; AVDictionaryEntry *tag; @@ -746,8 +820,10 @@ static int mkv_write_track(AVFormatContext *s, MatroskaMuxContext *mkv, return 0; } - if (!bit_depth) + if (!bit_depth && codec->codec_id != AV_CODEC_ID_ADPCM_G726) bit_depth = av_get_bytes_per_sample(codec->sample_fmt) << 3; + if (!bit_depth) + bit_depth = codec->bits_per_coded_sample; if (codec->codec_id == AV_CODEC_ID_AAC) { ret = get_aac_sample_rates(s, codec, &sample_rate, &output_sample_rate); @@ -756,59 +832,103 @@ static int mkv_write_track(AVFormatContext *s, MatroskaMuxContext *mkv, } track = start_ebml_master(pb, MATROSKA_ID_TRACKENTRY, 0); - put_ebml_uint (pb, MATROSKA_ID_TRACKNUMBER , i + 1); - put_ebml_uint (pb, MATROSKA_ID_TRACKUID , i + 1); + put_ebml_uint (pb, MATROSKA_ID_TRACKNUMBER, + mkv->is_dash ? mkv->dash_track_number : i + 1); + put_ebml_uint (pb, MATROSKA_ID_TRACKUID, + mkv->is_dash ? mkv->dash_track_number : i + 1); put_ebml_uint (pb, MATROSKA_ID_TRACKFLAGLACING , 0); // no lacing (yet) if ((tag = av_dict_get(st->metadata, "title", NULL, 0))) put_ebml_string(pb, MATROSKA_ID_TRACKNAME, tag->value); tag = av_dict_get(st->metadata, "language", NULL, 0); - put_ebml_string(pb, MATROSKA_ID_TRACKLANGUAGE, tag ? tag->value:"und"); + if (mkv->mode != MODE_WEBM || codec->codec_id != AV_CODEC_ID_WEBVTT) { + put_ebml_string(pb, MATROSKA_ID_TRACKLANGUAGE, tag && tag->value ? tag->value:"und"); + } else if (tag && tag->value) { + put_ebml_string(pb, MATROSKA_ID_TRACKLANGUAGE, tag->value); + } // The default value for TRACKFLAGDEFAULT is 1, so add element // if we need to clear it. - if (!(st->disposition & AV_DISPOSITION_DEFAULT)) + if (default_stream_exists && !(st->disposition & AV_DISPOSITION_DEFAULT)) put_ebml_uint(pb, MATROSKA_ID_TRACKFLAGDEFAULT, !!(st->disposition & AV_DISPOSITION_DEFAULT)); - if (codec->codec_type == AVMEDIA_TYPE_AUDIO && codec->initial_padding) { - mkv->tracks[i].ts_offset = av_rescale_q(codec->initial_padding, - (AVRational){ 1, codec->sample_rate }, - st->time_base); + if (st->disposition & AV_DISPOSITION_FORCED) + put_ebml_uint(pb, MATROSKA_ID_TRACKFLAGFORCED, 1); + + if (mkv->mode == MODE_WEBM && codec->codec_id == AV_CODEC_ID_WEBVTT) { + const char *codec_id; + if (st->disposition & AV_DISPOSITION_CAPTIONS) { + codec_id = "D_WEBVTT/CAPTIONS"; + native_id = MATROSKA_TRACK_TYPE_SUBTITLE; + } else if (st->disposition & AV_DISPOSITION_DESCRIPTIONS) { + codec_id = "D_WEBVTT/DESCRIPTIONS"; + native_id = MATROSKA_TRACK_TYPE_METADATA; + } else if (st->disposition & AV_DISPOSITION_METADATA) { + codec_id = "D_WEBVTT/METADATA"; + native_id = MATROSKA_TRACK_TYPE_METADATA; + } else { + codec_id = "D_WEBVTT/SUBTITLES"; + native_id = MATROSKA_TRACK_TYPE_SUBTITLE; + } + put_ebml_string(pb, MATROSKA_ID_CODECID, codec_id); + } else { + // look for a codec ID string specific to mkv to use, + // if none are found, use AVI codes + for (j = 0; ff_mkv_codec_tags[j].id != AV_CODEC_ID_NONE; j++) { + if (ff_mkv_codec_tags[j].id == codec->codec_id) { + put_ebml_string(pb, MATROSKA_ID_CODECID, ff_mkv_codec_tags[j].str); + native_id = 1; + break; + } + } + if (codec->codec_id == AV_CODEC_ID_RAWVIDEO && !codec->codec_tag) { + if (mkv->allow_raw_vfw) { + native_id = 0; + } else { + av_log(s, AV_LOG_ERROR, "Raw RGB is not supported Natively in Matroska, you can use AVI or NUT or\n" + "If you would like to store it anyway using VFW mode, enable allow_raw_vfw (-allow_raw_vfw 1)\n"); + return AVERROR(EINVAL); + } + } + } + + if (codec->codec_type == AVMEDIA_TYPE_AUDIO && codec->initial_padding && codec->codec_id == AV_CODEC_ID_OPUS) { +// mkv->tracks[i].ts_offset = av_rescale_q(codec->initial_padding, +// (AVRational){ 1, codec->sample_rate }, +// st->time_base); put_ebml_uint(pb, MATROSKA_ID_CODECDELAY, av_rescale_q(codec->initial_padding, (AVRational){ 1, codec->sample_rate }, (AVRational){ 1, 1000000000 })); } - - // look for a codec ID string specific to mkv to use, - // if none are found, use AVI codes - for (j = 0; ff_mkv_codec_tags[j].id != AV_CODEC_ID_NONE; j++) { - if (ff_mkv_codec_tags[j].id == codec->codec_id) { - put_ebml_string(pb, MATROSKA_ID_CODECID, ff_mkv_codec_tags[j].str); - native_id = 1; - break; - } + if (codec->codec_id == AV_CODEC_ID_OPUS) { + put_ebml_uint(pb, MATROSKA_ID_SEEKPREROLL, OPUS_SEEK_PREROLL); } if (mkv->mode == MODE_WEBM && !(codec->codec_id == AV_CODEC_ID_VP8 || codec->codec_id == AV_CODEC_ID_VP9 || codec->codec_id == AV_CODEC_ID_OPUS || - codec->codec_id == AV_CODEC_ID_VORBIS)) { + codec->codec_id == AV_CODEC_ID_VORBIS || + codec->codec_id == AV_CODEC_ID_WEBVTT)) { av_log(s, AV_LOG_ERROR, - "Only VP8 or VP9 video and Vorbis or Opus audio are supported for WebM.\n"); + "Only VP8 or VP9 video and Vorbis or Opus audio and WebVTT subtitles are supported for WebM.\n"); return AVERROR(EINVAL); } switch (codec->codec_type) { case AVMEDIA_TYPE_VIDEO: put_ebml_uint(pb, MATROSKA_ID_TRACKTYPE, MATROSKA_TRACK_TYPE_VIDEO); - if (st->avg_frame_rate.num > 0 && st->avg_frame_rate.den > 0) + + if( st->avg_frame_rate.num > 0 && st->avg_frame_rate.den > 0 + && 1.0/av_q2d(st->avg_frame_rate) > av_q2d(codec->time_base)) put_ebml_uint(pb, MATROSKA_ID_TRACKDEFAULTDURATION, 1E9 / av_q2d(st->avg_frame_rate)); + else + put_ebml_uint(pb, MATROSKA_ID_TRACKDEFAULTDURATION, av_q2d(codec->time_base)*1E9); if (!native_id && ff_codec_get_tag(ff_codec_movvideo_tags, codec->codec_id) && - (!ff_codec_get_tag(ff_codec_bmp_tags, codec->codec_id) || + ((!ff_codec_get_tag(ff_codec_bmp_tags, codec->codec_id) && codec->codec_id != AV_CODEC_ID_RAWVIDEO) || codec->codec_id == AV_CODEC_ID_SVQ1 || codec->codec_id == AV_CODEC_ID_SVQ3 || codec->codec_id == AV_CODEC_ID_CINEPAK)) @@ -829,10 +949,38 @@ static int mkv_write_track(AVFormatContext *s, MatroskaMuxContext *mkv, // check both side data and metadata for stereo information, // write the result to the bitstream if any is found - ret = mkv_write_stereo_mode(s, pb, st, mkv->mode); + ret = mkv_write_stereo_mode(s, pb, st, mkv->mode, + &display_width_div, + &display_height_div); if (ret < 0) return ret; + if (((tag = av_dict_get(st->metadata, "alpha_mode", NULL, 0)) && atoi(tag->value)) || + ((tag = av_dict_get( s->metadata, "alpha_mode", NULL, 0)) && atoi(tag->value)) || + (codec->pix_fmt == AV_PIX_FMT_YUVA420P)) { + put_ebml_uint(pb, MATROSKA_ID_VIDEOALPHAMODE, 1); + } + + // write DisplayWidth and DisplayHeight, they contain the size of + // a single source view and/or the display aspect ratio + if (st->sample_aspect_ratio.num) { + int64_t d_width = av_rescale(codec->width, st->sample_aspect_ratio.num, st->sample_aspect_ratio.den); + if (d_width > INT_MAX) { + av_log(s, AV_LOG_ERROR, "Overflow in display width\n"); + return AVERROR(EINVAL); + } + put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYWIDTH , d_width / display_width_div); + put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYHEIGHT, codec->height / display_height_div); + } else if (display_width_div != 1 || display_height_div != 1) { + put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYWIDTH , codec->width / display_width_div); + put_ebml_uint(pb, MATROSKA_ID_VIDEODISPLAYHEIGHT, codec->height / display_height_div); + } + + if (codec->codec_id == AV_CODEC_ID_RAWVIDEO) { + uint32_t color_space = av_le2ne32(codec->codec_tag); + put_ebml_binary(pb, MATROSKA_ID_VIDEOCOLORSPACE, &color_space, sizeof(color_space)); + } + end_ebml_master(pb, subinfo); break; @@ -854,19 +1002,26 @@ static int mkv_write_track(AVFormatContext *s, MatroskaMuxContext *mkv, break; case AVMEDIA_TYPE_SUBTITLE: - put_ebml_uint(pb, MATROSKA_ID_TRACKTYPE, MATROSKA_TRACK_TYPE_SUBTITLE); if (!native_id) { av_log(s, AV_LOG_ERROR, "Subtitle codec %d is not supported.\n", codec->codec_id); return AVERROR(ENOSYS); } + + if (mkv->mode != MODE_WEBM || codec->codec_id != AV_CODEC_ID_WEBVTT) + native_id = MATROSKA_TRACK_TYPE_SUBTITLE; + + put_ebml_uint(pb, MATROSKA_ID_TRACKTYPE, native_id); break; default: av_log(s, AV_LOG_ERROR, "Only audio, video, and subtitles are supported for Matroska.\n"); - break; + return AVERROR(EINVAL); + } + + if (mkv->mode != MODE_WEBM || codec->codec_id != AV_CODEC_ID_WEBVTT) { + ret = mkv_write_codecprivate(s, pb, codec, native_id, qt_id); + if (ret < 0) + return ret; } - ret = mkv_write_codecprivate(s, pb, codec, native_id, qt_id); - if (ret < 0) - return ret; end_ebml_master(pb, track); @@ -878,7 +1033,7 @@ static int mkv_write_tracks(AVFormatContext *s) MatroskaMuxContext *mkv = s->priv_data; AVIOContext *pb = s->pb; ebml_master tracks; - int i, ret; + int i, ret, default_stream_exists = 0; ret = mkv_add_seekhead_entry(mkv->main_seekhead, MATROSKA_ID_TRACKS, avio_tell(pb)); if (ret < 0) @@ -886,7 +1041,11 @@ static int mkv_write_tracks(AVFormatContext *s) tracks = start_ebml_master(pb, MATROSKA_ID_TRACKS, 0); for (i = 0; i < s->nb_streams; i++) { - ret = mkv_write_track(s, mkv, i, pb); + AVStream *st = s->streams[i]; + default_stream_exists |= st->disposition & AV_DISPOSITION_DEFAULT; + } + for (i = 0; i < s->nb_streams; i++) { + ret = mkv_write_track(s, mkv, i, pb, default_stream_exists); if (ret < 0) return ret; } @@ -922,7 +1081,7 @@ static int mkv_write_chapters(AVFormatContext *s) return AVERROR_INVALIDDATA; chapteratom = start_ebml_master(pb, MATROSKA_ID_CHAPTERATOM, 0); - put_ebml_uint(pb, MATROSKA_ID_CHAPTERUID, c->id); + put_ebml_uint(pb, MATROSKA_ID_CHAPTERUID, c->id + mkv->chapter_id_offset); put_ebml_uint(pb, MATROSKA_ID_CHAPTERTIMESTART, chapterstart); put_ebml_uint(pb, MATROSKA_ID_CHAPTERTIMEEND, chapterend); put_ebml_uint(pb, MATROSKA_ID_CHAPTERFLAGHIDDEN , 0); @@ -999,6 +1158,7 @@ static int mkv_write_tag(AVFormatContext *s, AVDictionary *m, unsigned int eleme while ((t = av_dict_get(m, "", t, AV_DICT_IGNORE_SUFFIX))) { if (av_strcasecmp(t->key, "title") && + av_strcasecmp(t->key, "stereo_mode") && av_strcasecmp(t->key, "encoding_tool")) { ret = mkv_write_simpletag(s->pb, t); if (ret < 0) @@ -1010,14 +1170,26 @@ static int mkv_write_tag(AVFormatContext *s, AVDictionary *m, unsigned int eleme return 0; } +static int mkv_check_tag(AVDictionary *m) +{ + AVDictionaryEntry *t = NULL; + + while ((t = av_dict_get(m, "", t, AV_DICT_IGNORE_SUFFIX))) + if (av_strcasecmp(t->key, "title") && av_strcasecmp(t->key, "stereo_mode")) + return 1; + + return 0; +} + static int mkv_write_tags(AVFormatContext *s) { + MatroskaMuxContext *mkv = s->priv_data; ebml_master tags = {0}; int i, ret; ff_metadata_conv_ctx(s, ff_mkv_metadata_conv, NULL); - if (av_dict_get(s->metadata, "", NULL, AV_DICT_IGNORE_SUFFIX)) { + if (mkv_check_tag(s->metadata)) { ret = mkv_write_tag(s, s->metadata, 0, 0, &tags); if (ret < 0) return ret; } @@ -1025,7 +1197,7 @@ static int mkv_write_tags(AVFormatContext *s) for (i = 0; i < s->nb_streams; i++) { AVStream *st = s->streams[i]; - if (!av_dict_get(st->metadata, "", 0, AV_DICT_IGNORE_SUFFIX)) + if (!mkv_check_tag(st->metadata)) continue; ret = mkv_write_tag(s, st->metadata, MATROSKA_ID_TAGTARGETS_TRACKUID, i + 1, &tags); @@ -1035,10 +1207,10 @@ static int mkv_write_tags(AVFormatContext *s) for (i = 0; i < s->nb_chapters; i++) { AVChapter *ch = s->chapters[i]; - if (!av_dict_get(ch->metadata, "", NULL, AV_DICT_IGNORE_SUFFIX)) + if (!mkv_check_tag(ch->metadata)) continue; - ret = mkv_write_tag(s, ch->metadata, MATROSKA_ID_TAGTARGETS_CHAPTERUID, ch->id, &tags); + ret = mkv_write_tag(s, ch->metadata, MATROSKA_ID_TAGTARGETS_CHAPTERUID, ch->id + mkv->chapter_id_offset, &tags); if (ret < 0) return ret; } @@ -1070,6 +1242,7 @@ static int mkv_write_attachments(AVFormatContext *s) ebml_master attached_file; AVDictionaryEntry *t; const char *mimetype = NULL; + uint64_t fileuid; if (st->codec->codec_type != AVMEDIA_TYPE_ATTACHMENT) continue; @@ -1099,9 +1272,25 @@ static int mkv_write_attachments(AVFormatContext *s) return AVERROR(EINVAL); } + if (s->flags & AVFMT_FLAG_BITEXACT) { + struct AVSHA *sha = av_sha_alloc(); + uint8_t digest[20]; + if (!sha) + return AVERROR(ENOMEM); + av_sha_init(sha, 160); + av_sha_update(sha, st->codec->extradata, st->codec->extradata_size); + av_sha_final(sha, digest); + av_free(sha); + fileuid = AV_RL64(digest); + } else { + fileuid = av_lfg_get(&c); + } + av_log(s, AV_LOG_VERBOSE, "Using %.16"PRIx64" for attachment %d\n", + fileuid, i); + put_ebml_string(pb, MATROSKA_ID_FILEMIMETYPE, mimetype); put_ebml_binary(pb, MATROSKA_ID_FILEDATA, st->codec->extradata, st->codec->extradata_size); - put_ebml_uint(pb, MATROSKA_ID_FILEUID, av_lfg_get(&c)); + put_ebml_uint(pb, MATROSKA_ID_FILEUID, fileuid); end_ebml_master(pb, attached_file); } end_ebml_master(pb, attachments); @@ -1115,14 +1304,40 @@ static int mkv_write_header(AVFormatContext *s) AVIOContext *pb = s->pb; ebml_master ebml_header, segment_info; AVDictionaryEntry *tag; - int ret, i; + int ret, i, version = 2; if (!strcmp(s->oformat->name, "webm")) mkv->mode = MODE_WEBM; else mkv->mode = MODE_MATROSKAv2; - mkv->tracks = av_mallocz(s->nb_streams * sizeof(*mkv->tracks)); + if (s->avoid_negative_ts < 0) + s->avoid_negative_ts = 1; + + if (mkv->mode != MODE_WEBM || + av_dict_get(s->metadata, "stereo_mode", NULL, 0) || + av_dict_get(s->metadata, "alpha_mode", NULL, 0)) + version = 4; + + for (i = 0; i < s->nb_streams; i++) { + if (s->streams[i]->codec->codec_id == AV_CODEC_ID_ATRAC3 || + s->streams[i]->codec->codec_id == AV_CODEC_ID_COOK || + s->streams[i]->codec->codec_id == AV_CODEC_ID_RA_288 || + s->streams[i]->codec->codec_id == AV_CODEC_ID_SIPR || + s->streams[i]->codec->codec_id == AV_CODEC_ID_RV10 || + s->streams[i]->codec->codec_id == AV_CODEC_ID_RV20) { + av_log(s, AV_LOG_ERROR, + "The Matroska muxer does not yet support muxing %s\n", + avcodec_get_name(s->streams[i]->codec->codec_id)); + return AVERROR_PATCHWELCOME; + } + if (s->streams[i]->codec->codec_id == AV_CODEC_ID_OPUS || + av_dict_get(s->streams[i]->metadata, "stereo_mode", NULL, 0) || + av_dict_get(s->streams[i]->metadata, "alpha_mode", NULL, 0)) + version = 4; + } + + mkv->tracks = av_mallocz_array(s->nb_streams, sizeof(*mkv->tracks)); if (!mkv->tracks) return AVERROR(ENOMEM); @@ -1132,7 +1347,7 @@ static int mkv_write_header(AVFormatContext *s) put_ebml_uint (pb, EBML_ID_EBMLMAXIDLENGTH , 4); put_ebml_uint (pb, EBML_ID_EBMLMAXSIZELENGTH , 8); put_ebml_string (pb, EBML_ID_DOCTYPE , s->oformat->name); - put_ebml_uint (pb, EBML_ID_DOCTYPEVERSION , 2); + put_ebml_uint (pb, EBML_ID_DOCTYPEVERSION , version); put_ebml_uint (pb, EBML_ID_DOCTYPEREADVERSION , 2); end_ebml_master(pb, ebml_header); @@ -1170,6 +1385,18 @@ static int mkv_write_header(AVFormatContext *s) else put_ebml_string(pb, MATROSKA_ID_WRITINGAPP, LIBAVFORMAT_IDENT); put_ebml_binary(pb, MATROSKA_ID_SEGMENTUID, segment_uid, 16); + } else { + const char *ident = "Lavf"; + put_ebml_string(pb, MATROSKA_ID_MUXINGAPP , ident); + put_ebml_string(pb, MATROSKA_ID_WRITINGAPP, ident); + } + + if (tag = av_dict_get(s->metadata, "creation_time", NULL, 0)) { + // Adjust time so it's relative to 2001-01-01 and convert to nanoseconds. + int64_t date_utc = (ff_iso8601_to_unix_time(tag->value) - 978307200) * 1000000000; + uint8_t date_utc_buf[8]; + AV_WB64(date_utc_buf, date_utc); + put_ebml_binary(pb, MATROSKA_ID_DATEUTC, date_utc_buf, 8); } // reserve space for the duration @@ -1182,6 +1409,9 @@ static int mkv_write_header(AVFormatContext *s) if (ret < 0) return ret; + for (i = 0; i < s->nb_chapters; i++) + mkv->chapter_id_offset = FFMAX(mkv->chapter_id_offset, 1LL - s->chapters[i]->id); + if (mkv->mode != MODE_WEBM) { ret = mkv_write_chapters(s); if (ret < 0) @@ -1210,6 +1440,7 @@ static int mkv_write_header(AVFormatContext *s) av_init_packet(&mkv->cur_audio_pkt); mkv->cur_audio_pkt.size = 0; + mkv->cluster_pos = -1; avio_flush(pb); @@ -1241,85 +1472,6 @@ static int mkv_blockgroup_size(int pkt_size) return size; } -static int ass_get_duration(AVFormatContext *s, const uint8_t *p) -{ - int sh, sm, ss, sc, eh, em, es, ec; - uint64_t start, end; - - if (sscanf(p, "%*[^,],%d:%d:%d%*c%d,%d:%d:%d%*c%d", - &sh, &sm, &ss, &sc, &eh, &em, &es, &ec) != 8) - return 0; - - if (sh > 9 || sm > 59 || ss > 59 || sc > 99 || - eh > 9 || em > 59 || es > 59 || ec > 99) { - av_log(s, AV_LOG_WARNING, - "Non-standard time reference %d:%d:%d.%d,%d:%d:%d.%d\n", - sh, sm, ss, sc, eh, em, es, ec); - return 0; - } - - start = 3600000 * sh + 60000 * sm + 1000 * ss + 10 * sc; - end = 3600000 * eh + 60000 * em + 1000 * es + 10 * ec; - - if (start > end) { - av_log(s, AV_LOG_WARNING, - "Unexpected time reference %d:%d:%d.%d,%d:%d:%d.%d\n", - sh, sm, ss, sc, eh, em, es, ec); - return 0; - } - - return end - start; -} - -static int mkv_write_ass_blocks(AVFormatContext *s, AVIOContext *pb, - AVPacket *pkt) -{ - MatroskaMuxContext *mkv = s->priv_data; - int i, layer = 0, max_duration = 0, size, line_size, data_size = pkt->size; - uint8_t *start, *end, *data = pkt->data; - ebml_master blockgroup; - char buffer[2048]; - - while (data_size) { - int duration = ass_get_duration(s, data); - max_duration = FFMAX(duration, max_duration); - end = memchr(data, '\n', data_size); - size = line_size = end ? end - data + 1 : data_size; - size -= end ? (end[-1] == '\r') + 1 : 0; - start = data; - for (i = 0; i < 3; i++, start++) - if (!(start = memchr(start, ',', size - (start - data)))) - return max_duration; - size -= start - data; - sscanf(data, "Dialogue: %d,", &layer); - i = snprintf(buffer, sizeof(buffer), "%" PRId64 ",%d,", - s->streams[pkt->stream_index]->nb_frames, layer); - size = FFMIN(i + size, sizeof(buffer)); - memcpy(buffer + i, start, size - i); - - av_log(s, AV_LOG_DEBUG, - "Writing block at offset %" PRIu64 ", size %d, " - "pts %" PRId64 ", duration %d\n", - avio_tell(pb), size, pkt->pts, duration); - blockgroup = start_ebml_master(pb, MATROSKA_ID_BLOCKGROUP, - mkv_blockgroup_size(size)); - put_ebml_id(pb, MATROSKA_ID_BLOCK); - put_ebml_num(pb, size + 4, 0); - // this assumes stream_index is less than 126 - avio_w8(pb, 0x80 | (pkt->stream_index + 1)); - avio_wb16(pb, pkt->pts - mkv->cluster_pts); - avio_w8(pb, 0); - avio_write(pb, buffer, size); - put_ebml_uint(pb, MATROSKA_ID_BLOCKDURATION, duration); - end_ebml_master(pb, blockgroup); - - data += line_size; - data_size -= line_size; - } - - return max_duration; -} - static int mkv_strip_wavpack(const uint8_t *src, uint8_t **pdst, int *size) { uint8_t *dst; @@ -1378,9 +1530,12 @@ static void mkv_write_block(AVFormatContext *s, AVIOContext *pb, { MatroskaMuxContext *mkv = s->priv_data; AVCodecContext *codec = s->streams[pkt->stream_index]->codec; - uint8_t *data = NULL; - int offset = 0, size = pkt->size; + uint8_t *data = NULL, *side_data = NULL; + int offset = 0, size = pkt->size, side_data_size = 0; int64_t ts = mkv->tracks[pkt->stream_index].write_dts ? pkt->dts : pkt->pts; + uint64_t additional_id = 0; + int64_t discard_padding = 0; + ebml_master block_group, block_additions, block_more; av_log(s, AV_LOG_DEBUG, "Writing block at offset %" PRIu64 ", size %d, " "pts %" PRId64 ", dts %" PRId64 ", duration %d, flags %d\n", @@ -1408,56 +1563,95 @@ static void mkv_write_block(AVFormatContext *s, AVIOContext *pb, offset = 8; } + side_data = av_packet_get_side_data(pkt, + AV_PKT_DATA_SKIP_SAMPLES, + &side_data_size); + + if (side_data && side_data_size >= 10) { + discard_padding = av_rescale_q(AV_RL32(side_data + 4), + (AVRational){1, codec->sample_rate}, + (AVRational){1, 1000000000}); + } + + side_data = av_packet_get_side_data(pkt, + AV_PKT_DATA_MATROSKA_BLOCKADDITIONAL, + &side_data_size); + if (side_data) { + additional_id = AV_RB64(side_data); + side_data += 8; + side_data_size -= 8; + } + + if ((side_data_size && additional_id == 1) || discard_padding) { + block_group = start_ebml_master(pb, MATROSKA_ID_BLOCKGROUP, 0); + blockid = MATROSKA_ID_BLOCK; + } + put_ebml_id(pb, blockid); put_ebml_num(pb, size + 4, 0); // this assumes stream_index is less than 126 - avio_w8(pb, 0x80 | (pkt->stream_index + 1)); + avio_w8(pb, 0x80 | (mkv->is_dash ? mkv->dash_track_number : (pkt->stream_index + 1))); avio_wb16(pb, ts - mkv->cluster_pts); avio_w8(pb, flags); avio_write(pb, data + offset, size); if (data != pkt->data) av_free(data); -} -static int srt_get_duration(uint8_t **buf) -{ - int i, duration = 0; - - for (i = 0; i < 2 && !duration; i++) { - int s_hour, s_min, s_sec, s_hsec, e_hour, e_min, e_sec, e_hsec; - if (sscanf(*buf, "%d:%2d:%2d%*1[,.]%3d --> %d:%2d:%2d%*1[,.]%3d", - &s_hour, &s_min, &s_sec, &s_hsec, - &e_hour, &e_min, &e_sec, &e_hsec) == 8) { - s_min += 60 * s_hour; - e_min += 60 * e_hour; - s_sec += 60 * s_min; - - e_sec += 60 * e_min; - s_hsec += 1000 * s_sec; - e_hsec += 1000 * e_sec; - - duration = e_hsec - s_hsec; - } - *buf += strcspn(*buf, "\n") + 1; + if (discard_padding) { + put_ebml_sint(pb, MATROSKA_ID_DISCARDPADDING, discard_padding); + } + + if (side_data_size && additional_id == 1) { + block_additions = start_ebml_master(pb, MATROSKA_ID_BLOCKADDITIONS, 0); + block_more = start_ebml_master(pb, MATROSKA_ID_BLOCKMORE, 0); + put_ebml_uint(pb, MATROSKA_ID_BLOCKADDID, 1); + put_ebml_id(pb, MATROSKA_ID_BLOCKADDITIONAL); + put_ebml_num(pb, side_data_size, 0); + avio_write(pb, side_data, side_data_size); + end_ebml_master(pb, block_more); + end_ebml_master(pb, block_additions); + } + if ((side_data_size && additional_id == 1) || discard_padding) { + end_ebml_master(pb, block_group); } - return duration; } -static int mkv_write_srt_blocks(AVFormatContext *s, AVIOContext *pb, - AVPacket *pkt) +static int mkv_write_vtt_blocks(AVFormatContext *s, AVIOContext *pb, AVPacket *pkt) { + MatroskaMuxContext *mkv = s->priv_data; ebml_master blockgroup; - AVPacket pkt2 = *pkt; - int64_t duration = srt_get_duration(&pkt2.data); - pkt2.size -= pkt2.data - pkt->data; - - blockgroup = start_ebml_master(pb, MATROSKA_ID_BLOCKGROUP, - mkv_blockgroup_size(pkt2.size)); - mkv_write_block(s, pb, MATROSKA_ID_BLOCK, &pkt2, 0); - put_ebml_uint(pb, MATROSKA_ID_BLOCKDURATION, duration); + int id_size, settings_size, size; + uint8_t *id, *settings; + int64_t ts = mkv->tracks[pkt->stream_index].write_dts ? pkt->dts : pkt->pts; + const int flags = 0; + + id_size = 0; + id = av_packet_get_side_data(pkt, AV_PKT_DATA_WEBVTT_IDENTIFIER, + &id_size); + + settings_size = 0; + settings = av_packet_get_side_data(pkt, AV_PKT_DATA_WEBVTT_SETTINGS, + &settings_size); + + size = id_size + 1 + settings_size + 1 + pkt->size; + + av_log(s, AV_LOG_DEBUG, "Writing block at offset %" PRIu64 ", size %d, " + "pts %" PRId64 ", dts %" PRId64 ", duration %d, flags %d\n", + avio_tell(pb), size, pkt->pts, pkt->dts, pkt->duration, flags); + + blockgroup = start_ebml_master(pb, MATROSKA_ID_BLOCKGROUP, mkv_blockgroup_size(size)); + + put_ebml_id(pb, MATROSKA_ID_BLOCK); + put_ebml_num(pb, size + 4, 0); + avio_w8(pb, 0x80 | (pkt->stream_index + 1)); // this assumes stream_index is less than 126 + avio_wb16(pb, ts - mkv->cluster_pts); + avio_w8(pb, flags); + avio_printf(pb, "%.*s\n%.*s\n%.*s", id_size, id, settings_size, settings, pkt->size, pkt->data); + + put_ebml_uint(pb, MATROSKA_ID_BLOCKDURATION, pkt->duration); end_ebml_master(pb, blockgroup); - return duration; + return pkt->duration; } static void mkv_flush_dynbuf(AVFormatContext *s) @@ -1475,7 +1669,29 @@ static void mkv_flush_dynbuf(AVFormatContext *s) mkv->dyn_bc = NULL; } -static int mkv_write_packet_internal(AVFormatContext *s, AVPacket *pkt) +static void mkv_start_new_cluster(AVFormatContext *s, AVPacket *pkt) +{ + MatroskaMuxContext *mkv = s->priv_data; + AVIOContext *pb; + + if (s->pb->seekable) { + pb = s->pb; + } else { + pb = mkv->dyn_bc; + } + + av_log(s, AV_LOG_DEBUG, + "Starting new cluster at offset %" PRIu64 " bytes, " + "pts %" PRIu64 "dts %" PRIu64 "\n", + avio_tell(pb), pkt->pts, pkt->dts); + end_ebml_master(pb, mkv->cluster); + mkv->cluster_pos = -1; + if (mkv->dyn_bc) + mkv_flush_dynbuf(s); + avio_flush(s->pb); +} + +static int mkv_write_packet_internal(AVFormatContext *s, AVPacket *pkt, int add_cue) { MatroskaMuxContext *mkv = s->priv_data; AVIOContext *pb = s->pb; @@ -1484,6 +1700,8 @@ static int mkv_write_packet_internal(AVFormatContext *s, AVPacket *pkt) int duration = pkt->duration; int ret; int64_t ts = mkv->tracks[pkt->stream_index].write_dts ? pkt->dts : pkt->pts; + int64_t relative_packet_pos; + int dash_tracknum = mkv->is_dash ? mkv->dash_track_number : pkt->stream_index + 1; if (ts == AV_NOPTS_VALUE) { av_log(s, AV_LOG_ERROR, "Can't write packet with unknown timestamp\n"); @@ -1491,42 +1709,59 @@ static int mkv_write_packet_internal(AVFormatContext *s, AVPacket *pkt) } ts += mkv->tracks[pkt->stream_index].ts_offset; + if (mkv->cluster_pos != -1) { + int64_t cluster_time = ts - mkv->cluster_pts + mkv->tracks[pkt->stream_index].ts_offset; + if ((int16_t)cluster_time != cluster_time) { + av_log(s, AV_LOG_WARNING, "Starting new cluster due to timestamp\n"); + mkv_start_new_cluster(s, pkt); + } + } + if (!s->pb->seekable) { if (!mkv->dyn_bc) { ret = avio_open_dyn_buf(&mkv->dyn_bc); - if (ret < 0) + if (ret < 0) { + av_log(s, AV_LOG_ERROR, "Failed to open dynamic buffer\n"); return ret; + } } pb = mkv->dyn_bc; } - if (!mkv->cluster_pos) { + if (mkv->cluster_pos == -1) { mkv->cluster_pos = avio_tell(s->pb); mkv->cluster = start_ebml_master(pb, MATROSKA_ID_CLUSTER, 0); put_ebml_uint(pb, MATROSKA_ID_CLUSTERTIMECODE, FFMAX(0, ts)); mkv->cluster_pts = FFMAX(0, ts); } + relative_packet_pos = avio_tell(s->pb) - mkv->cluster.pos; + if (codec->codec_type != AVMEDIA_TYPE_SUBTITLE) { mkv_write_block(s, pb, MATROSKA_ID_SIMPLEBLOCK, pkt, keyframe << 7); - } else if (codec->codec_id == AV_CODEC_ID_SSA) { - duration = mkv_write_ass_blocks(s, pb, pkt); - } else if (codec->codec_id == AV_CODEC_ID_SRT) { - duration = mkv_write_srt_blocks(s, pb, pkt); + if (codec->codec_type == AVMEDIA_TYPE_VIDEO && keyframe || add_cue) { + ret = mkv_add_cuepoint(mkv->cues, pkt->stream_index, dash_tracknum, ts, mkv->cluster_pos, relative_packet_pos, -1); + if (ret < 0) return ret; + } + } else { + if (codec->codec_id == AV_CODEC_ID_WEBVTT) { + duration = mkv_write_vtt_blocks(s, pb, pkt); } else { ebml_master blockgroup = start_ebml_master(pb, MATROSKA_ID_BLOCKGROUP, mkv_blockgroup_size(pkt->size)); - duration = pkt->convergence_duration; + /* For backward compatibility, prefer convergence_duration. */ + if (pkt->convergence_duration > 0) { + duration = pkt->convergence_duration; + } mkv_write_block(s, pb, MATROSKA_ID_BLOCK, pkt, 0); put_ebml_uint(pb, MATROSKA_ID_BLOCKDURATION, duration); end_ebml_master(pb, blockgroup); } - if (codec->codec_type == AVMEDIA_TYPE_VIDEO && keyframe) { - ret = mkv_add_cuepoint(mkv->cues, pkt->stream_index, ts, - mkv->cluster_pos); - if (ret < 0) - return ret; + ret = mkv_add_cuepoint(mkv->cues, pkt->stream_index, dash_tracknum, ts, + mkv->cluster_pos, relative_packet_pos, duration); + if (ret < 0) + return ret; } mkv->duration = FFMAX(mkv->duration, ts + duration); @@ -1540,8 +1775,8 @@ static int mkv_write_packet(AVFormatContext *s, AVPacket *pkt) int keyframe = !!(pkt->flags & AV_PKT_FLAG_KEY); int cluster_size; int64_t cluster_time; - AVIOContext *pb; int ret; + int start_new_cluster; if (mkv->tracks[pkt->stream_index].write_dts) cluster_time = pkt->dts - mkv->cluster_pts; @@ -1552,32 +1787,40 @@ static int mkv_write_packet(AVFormatContext *s, AVPacket *pkt) // start a new cluster every 5 MB or 5 sec, or 32k / 1 sec for streaming or // after 4k and on a keyframe if (s->pb->seekable) { - pb = s->pb; - cluster_size = avio_tell(pb) - mkv->cluster_pos; + cluster_size = avio_tell(s->pb) - mkv->cluster_pos; } else { - pb = mkv->dyn_bc; - cluster_size = avio_tell(pb); - } - - if (mkv->cluster_pos && - (cluster_size > mkv->cluster_size_limit || - cluster_time > mkv->cluster_time_limit || - (codec_type == AVMEDIA_TYPE_VIDEO && keyframe && - cluster_size > 4 * 1024))) { - av_log(s, AV_LOG_DEBUG, - "Starting new cluster at offset %" PRIu64 " bytes, " - "pts %" PRIu64 "dts %" PRIu64 "\n", - avio_tell(pb), pkt->pts, pkt->dts); - end_ebml_master(pb, mkv->cluster); - mkv->cluster_pos = 0; - if (mkv->dyn_bc) - mkv_flush_dynbuf(s); - avio_flush(s->pb); + cluster_size = avio_tell(mkv->dyn_bc); + } + + if (mkv->is_dash && codec_type == AVMEDIA_TYPE_VIDEO) { + // WebM DASH specification states that the first block of every cluster + // has to be a key frame. So for DASH video, we only create a cluster + // on seeing key frames. + start_new_cluster = keyframe; + } else if (mkv->is_dash && codec_type == AVMEDIA_TYPE_AUDIO && + (mkv->cluster_pos == -1 || + cluster_time > mkv->cluster_time_limit)) { + // For DASH audio, we create a Cluster based on cluster_time_limit + start_new_cluster = 1; + } else if (!mkv->is_dash && + (cluster_size > mkv->cluster_size_limit || + cluster_time > mkv->cluster_time_limit || + (codec_type == AVMEDIA_TYPE_VIDEO && keyframe && + cluster_size > 4 * 1024))) { + start_new_cluster = 1; + } else { + start_new_cluster = 0; + } + + if (mkv->cluster_pos != -1 && start_new_cluster) { + mkv_start_new_cluster(s, pkt); } // check if we have an audio packet cached if (mkv->cur_audio_pkt.size > 0) { - ret = mkv_write_packet_internal(s, &mkv->cur_audio_pkt); + // for DASH audio, a CuePoint has to be added when there is a new cluster. + ret = mkv_write_packet_internal(s, &mkv->cur_audio_pkt, + mkv->is_dash ? start_new_cluster : 0); av_free_packet(&mkv->cur_audio_pkt); if (ret < 0) { av_log(s, AV_LOG_ERROR, @@ -1595,8 +1838,11 @@ static int mkv_write_packet(AVFormatContext *s, AVPacket *pkt) ret = mkv->cur_audio_pkt.buf ? 0 : AVERROR(ENOMEM); } else ret = av_dup_packet(&mkv->cur_audio_pkt); + if (mkv->cur_audio_pkt.side_data_elems > 0) { + ret = av_copy_packet_side_data(&mkv->cur_audio_pkt, &mkv->cur_audio_pkt); + } } else - ret = mkv_write_packet_internal(s, pkt); + ret = mkv_write_packet_internal(s, pkt, 0); return ret; } @@ -1609,12 +1855,12 @@ static int mkv_write_flush_packet(AVFormatContext *s, AVPacket *pkt) else pb = mkv->dyn_bc; if (!pkt) { - if (mkv->cluster_pos) { + if (mkv->cluster_pos != -1) { av_log(s, AV_LOG_DEBUG, "Flushing cluster at offset %" PRIu64 " bytes\n", avio_tell(pb)); end_ebml_master(pb, mkv->cluster); - mkv->cluster_pos = 0; + mkv->cluster_pos = -1; if (mkv->dyn_bc) mkv_flush_dynbuf(s); avio_flush(s->pb); @@ -1633,7 +1879,7 @@ static int mkv_write_trailer(AVFormatContext *s) // check if we have an audio packet cached if (mkv->cur_audio_pkt.size > 0) { - ret = mkv_write_packet_internal(s, &mkv->cur_audio_pkt); + ret = mkv_write_packet_internal(s, &mkv->cur_audio_pkt, 0); av_free_packet(&mkv->cur_audio_pkt); if (ret < 0) { av_log(s, AV_LOG_ERROR, @@ -1645,7 +1891,7 @@ static int mkv_write_trailer(AVFormatContext *s) if (mkv->dyn_bc) { end_ebml_master(mkv->dyn_bc, mkv->cluster); mkv_flush_dynbuf(s); - } else if (mkv->cluster_pos) { + } else if (mkv->cluster_pos != -1) { end_ebml_master(pb, mkv->cluster); } @@ -1663,7 +1909,7 @@ static int mkv_write_trailer(AVFormatContext *s) currentpos = avio_tell(pb); avio_seek(pb, mkv->cues_pos, SEEK_SET); - cuespos = mkv_write_cues(pb, mkv->cues, s->nb_streams); + cuespos = mkv_write_cues(s, mkv->cues, mkv->tracks, s->nb_streams); cues_end = avio_tell(pb); if (cues_end > cuespos + mkv->reserve_cues_space) { av_log(s, AV_LOG_ERROR, @@ -1679,7 +1925,7 @@ static int mkv_write_trailer(AVFormatContext *s) avio_seek(pb, currentpos, SEEK_SET); } else { - cuespos = mkv_write_cues(pb, mkv->cues, s->nb_streams); + cuespos = mkv_write_cues(s, mkv->cues, mkv->tracks, s->nb_streams); } ret = mkv_add_seekhead_entry(mkv->main_seekhead, MATROSKA_ID_CUES, @@ -1700,7 +1946,7 @@ static int mkv_write_trailer(AVFormatContext *s) } end_ebml_master(pb, mkv->segment); - av_free(mkv->tracks); + av_freep(&mkv->tracks); av_freep(&mkv->cues->entries); av_freep(&mkv->cues); @@ -1724,12 +1970,40 @@ static int mkv_query_codec(enum AVCodecID codec_id, int std_compliance) return 0; } +static const AVCodecTag additional_audio_tags[] = { + { AV_CODEC_ID_ALAC, 0XFFFFFFFF }, + { AV_CODEC_ID_EAC3, 0XFFFFFFFF }, + { AV_CODEC_ID_MLP, 0xFFFFFFFF }, + { AV_CODEC_ID_OPUS, 0xFFFFFFFF }, + { AV_CODEC_ID_PCM_S16BE, 0xFFFFFFFF }, + { AV_CODEC_ID_PCM_S24BE, 0xFFFFFFFF }, + { AV_CODEC_ID_PCM_S32BE, 0xFFFFFFFF }, + { AV_CODEC_ID_QDM2, 0xFFFFFFFF }, + { AV_CODEC_ID_RA_144, 0xFFFFFFFF }, + { AV_CODEC_ID_RA_288, 0xFFFFFFFF }, + { AV_CODEC_ID_COOK, 0xFFFFFFFF }, + { AV_CODEC_ID_TRUEHD, 0xFFFFFFFF }, + { AV_CODEC_ID_NONE, 0xFFFFFFFF } +}; + +static const AVCodecTag additional_video_tags[] = { + { AV_CODEC_ID_RV10, 0xFFFFFFFF }, + { AV_CODEC_ID_RV20, 0xFFFFFFFF }, + { AV_CODEC_ID_RV30, 0xFFFFFFFF }, + { AV_CODEC_ID_RV40, 0xFFFFFFFF }, + { AV_CODEC_ID_VP9, 0xFFFFFFFF }, + { AV_CODEC_ID_NONE, 0xFFFFFFFF } +}; + #define OFFSET(x) offsetof(MatroskaMuxContext, x) #define FLAGS AV_OPT_FLAG_ENCODING_PARAM static const AVOption options[] = { { "reserve_index_space", "Reserve a given amount of space (in bytes) at the beginning of the file for the index (cues).", OFFSET(reserve_cues_space), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS }, { "cluster_size_limit", "Store at most the provided amount of bytes in a cluster. ", OFFSET(cluster_size_limit), AV_OPT_TYPE_INT , { .i64 = -1 }, -1, INT_MAX, FLAGS }, { "cluster_time_limit", "Store at most the provided number of milliseconds in a cluster.", OFFSET(cluster_time_limit), AV_OPT_TYPE_INT64, { .i64 = -1 }, -1, INT64_MAX, FLAGS }, + { "dash", "Create a WebM file conforming to WebM DASH specification", OFFSET(is_dash), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, FLAGS }, + { "dash_track_number", "Track number for the DASH stream", OFFSET(dash_track_number), AV_OPT_TYPE_INT, { .i64 = 1 }, 0, 127, FLAGS }, + { "allow_raw_vfw", "allow RAW VFW mode", OFFSET(allow_raw_vfw), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, FLAGS }, { NULL }, }; @@ -1757,9 +2031,10 @@ AVOutputFormat ff_matroska_muxer = { .flags = AVFMT_GLOBALHEADER | AVFMT_VARIABLE_FPS | AVFMT_TS_NONSTRICT | AVFMT_ALLOW_FLUSH, .codec_tag = (const AVCodecTag* const []){ - ff_codec_bmp_tags, ff_codec_wav_tags, 0 + ff_codec_bmp_tags, ff_codec_wav_tags, + additional_audio_tags, additional_video_tags, 0 }, - .subtitle_codec = AV_CODEC_ID_SSA, + .subtitle_codec = AV_CODEC_ID_ASS, .query_codec = mkv_query_codec, .priv_class = &matroska_class, }; @@ -1781,6 +2056,7 @@ AVOutputFormat ff_webm_muxer = { .priv_data_size = sizeof(MatroskaMuxContext), .audio_codec = AV_CODEC_ID_VORBIS, .video_codec = AV_CODEC_ID_VP8, + .subtitle_codec = AV_CODEC_ID_WEBVTT, .write_header = mkv_write_header, .write_packet = mkv_write_flush_packet, .write_trailer = mkv_write_trailer, @@ -1799,7 +2075,7 @@ static const AVClass mka_class = { }; AVOutputFormat ff_matroska_audio_muxer = { .name = "matroska", - .long_name = NULL_IF_CONFIG_SMALL("Matroska"), + .long_name = NULL_IF_CONFIG_SMALL("Matroska Audio"), .mime_type = "audio/x-matroska", .extensions = "mka", .priv_data_size = sizeof(MatroskaMuxContext), @@ -1811,7 +2087,9 @@ AVOutputFormat ff_matroska_audio_muxer = { .write_trailer = mkv_write_trailer, .flags = AVFMT_GLOBALHEADER | AVFMT_TS_NONSTRICT | AVFMT_ALLOW_FLUSH, - .codec_tag = (const AVCodecTag* const []){ ff_codec_wav_tags, 0 }, + .codec_tag = (const AVCodecTag* const []){ + ff_codec_wav_tags, additional_audio_tags, 0 + }, .priv_class = &mka_class, }; #endif |