1 Implementing GstToc support in GStreamer elements
3 1. General info about GstToc structure
5 GstToc introduces a general way to handle chapters within multimedia formats.
6 GstToc can be represented as tree structure with arbitrary hierarchy. Tree item
7 can be either of two types: sequence or alternative. Sequence types acts like a
8 part of the media data, for example audio track in CUE sheet, or part of the
9 movie. Alternative types acts like some kind of selection to process a different
10 version of the media content, for example DVD angles.
11 GstToc has one constraint on the tree structure: it does not allow different
12 entry types on the same level of the hierarchy, i.e. you shouldn't
13 have editions and chapters mixed together. Here is an example of right TOC:
22 Here are two editions (alternatives), the first contains two chapters (sequence
23 type), and the second has only one chapter. And here is an example of invalid
33 Here you have edition1 and chapter1 mixed on the same level of hierarchy,
34 and such TOC will be considered broken.
36 GstToc has 'entries' field of GList type which consists of children items.
37 Each item is of type GstTocEntry. Also GstToc has list of tags and
38 GstStructure called 'info'. Please, use GstToc.info and GstTocEntry.info
39 fields this way: create a GstStructure, put all info related to your element
40 there and put this structure into the 'info' field under the name of your
41 element. Some fields in the 'info' structure can be used for internal
42 purposes, so you should use it in the way described above to not to
43 overwrite already existent fields.
45 Let's look at GstTocEntry a bit closer. One of the most important fields
46 is 'uid', which must be unique for each item within the TOC. This is used
47 to identify each item inside TOC, especially when element receives TOC
48 select event with UID to seek on. Field 'subentries' of type GList contains
49 children items of type GstTocEntry. Thus you can achieve arbitrary hierarchy
50 level. Field 'type' can be either GST_TOC_ENTRY_TYPE_CHAPTER or
51 GST_TOC_ENTRY_TYPE_EDITION which corresponds to chapter or edition type of
52 item respectively. Field 'tags' is a list of tags related to the item. And field
53 'info' is similar to GstToc.info described above.
55 So, a little more about managing GstToc. Use gst_toc_new() and gst_toc_unref()
56 to create/free it. GstTocEntry can be created using gst_toc_entry_new().
57 While building GstToc you can set start and stop timestamps for each item using
58 gst_toc_entry_set_start_stop() and loop_type and repeat_count using
59 gst_toc_entry_set_loop().
60 The best way to process already created GstToc is to recursively go through
61 the 'entries' and 'subentries' fields.
63 Applications and plugins should not rely on TOCs having a certain kind of
64 structure, but should allow for different alternatives. For example, a
65 simple CUE sheet embedded in a file may be presented as a flat list of
66 track entries, or could have a top-level edition node (or some other
67 alternative type entry) with track entries underneath that node; or even
68 multiple top-level edition nodes (or some other alternative type entries)
69 each with track entries underneath, in case the source file has extracted
70 a track listing from different sources).
73 2. TOC scope: global and current
75 There are two main consumers for TOC information: applications and elements
76 in the pipeline that are TOC writers (such as e.g. matroskamux).
78 Applications typically want to know the entire table of contents (TOC) with
79 all entries that can possibly be selected.
81 TOC writers in the pipeline, however, would not want to write a TOC for all
82 possible/available streams, but only for the current stream.
84 When transcoding a title from a DVD, for example, the application would still
85 want to know the entire TOC, with all titles, the chapters for each title,
86 and the available angles. When transcoding to a file, we only want the TOC
87 information that is relevant to the transcoded stream to be written into
88 the file structure, e.g. the chapters of the title being transcoded (or
89 possibly only chapters 5-7 if only those have been selected for playback/
92 This is why we may need to create two different TOCs for those two types
95 Elements that extract TOC information should send TOC events downstream.
97 Like with tags, sinks will post a TOC message on the bus for the application
98 with the global TOC, once a global TOC event reaches the sink.
101 3. Working with GstMessage
103 If a table of contents is available, applications will receive a TOC message
104 on the pipeline's GstBus.
106 A TOC message will be posted on the bus by sinks when the receive a TOC event
107 containing a TOC with global scope. Elements extracting TOCs should not post
108 a TOC message themselves, but send a TOC event downstream.
110 The reason for this is that there may be cascades of TOCs (e.g. a zip archive
111 containing multiple matroska files, each with a TOC).
113 GstMessage with GstToc can be created using gst_message_new_toc() and parsed
114 with gst_message_parse_toc(). The 'updated' parameter in these methods indicates
115 whether the TOC was just discovered (set to false) or TOC was already found and
116 have been updated (set to true). This message will typically be posted by sinks
117 to pipeline in case you have discovered TOC data within your element.
120 4. Working with GstEvent
122 There are two types of TOC-related events:
124 - downstream TOC events that contain TOC information and
127 - toc-select events that travel upstream and can be used to select
128 a certain TOC entry for playback (similar to seek events)
130 GstToc supports select event through GstEvent infrastructure. The idea is the
131 following: when you receive TOC select event, parse it with
132 gst_event_parse_toc_select() and seek stream (if it is not streamable) for
133 specified TOC UID (you can use gst_toc_find_entry() to find entry in TOC by UID).
134 To create TOC select event use gst_event_new_toc_select(). The common action on
135 such event is to seek to specified UID within your element.
138 5. Implementation coverage, Specifications, ...
140 Below is a list of container formats, links to documentation and a summary of
141 toc related features. Each section title also indicates whether reading/writing
142 a toc is implemented. Below hollow bullet point 'o' indicate no support and
143 filled bullets '*' indicate that this feature is handled.
147 http://www-mmsp.ece.mcgill.ca/Documents/AudioFormats/AIFF/Docs/AIFF-1.3.pdf
151 The 'MARK' chunk defines a list of (cue-id, position_in_samples, label).
153 The 'INST' chunk contains a sustainLoop and releaseLoop, each consisting of
154 (loop-type, cue-begin, cue-end)
158 http://xiph.org/flac/format.html#metadata_block_cuesheet
159 * METADATA_BLOCK_CUESHEET
161 o CUESHEET_TRACK_INDEX
163 Both CUESHEET_TRACK and CUESHEET_TRACK_INDEX have a (relative) offset in
164 samples. CUESHEET_TRACK has ISRC metadata.
168 http://matroska.org/technical/specs/chapters/index.html
169 * Chapters and Editions each having a uid
170 * Chapter have start/end time and metadata:
171 ChapString, ChapLanguage, ChapCountry
177 The 'elst' atom contains a list of edits. Each edit consists of (length, start,
182 https://wiki.xiph.org/Chapter_Extension
183 o VorbisComment fields called CHAPTERxxx and CHAPTERxxxNAME with xxx being a
184 number between 000 and 999.
188 http://www.sonicspot.com/guide/wavefiles.html
197 The 'cue ' chunk defines a list of markers in the stream with 'cue-id's. The
198 'smpl' chunk defines a list of regions in the stream with 'cue-id's in the same
201 The various 'adtl' chunks: 'labl', 'note' and 'ltxt' refer to the 'cue-id's.
203 A 'plst' chunk defines a sequence of segments (cue-id, length_samples, repeats).
204 The 'smpl' chunk defines a list of loops (cue-id, beg, end, loop-type, repeats).
207 6. Conclusion/Ideas/Future work
209 Based on the data of chapter 5, a few thoughts and observations that can be used
210 to extend and refine our API. These things below are not reflecting the current
213 All formats have table of [cue-id, cue-start, (cue-end), (extra tags)]
214 - cue-id is commonly represented as and unsigned int 32bit
215 - cue-end is optional
216 - extra tags could be represented as a structure/taglist
218 Many formats have metadata that references the cue-table.
219 - loops in instruments in wav, aifc
220 - edit lists in wav, mp4
222 For mp4.edtl, wav.plst we could expose two editions.
223 1) the edit list is flattened: default, for playback
224 2) the stream has the raw data and the edit list is there as chapter markers:
225 useful for editing software
227 We might want to introduce a new GST_TOC_ENTRY_TYPE_MARKER or _CUE. This would
228 be a sequence entry-type and it would not be used for navigational purposes, but
229 to attach data to a point in time (envelopes, loops, ...).
231 API wise there is some overlap between:
232 - exposing multiple audio/video tracks as pads or as ToC editions. For ToC
233 editions, we have the TocSelect event.
234 - exposing subtitles as a sparse stream or as as ToC sequence of markers with