Re: [Cellar] when is ChapterEndTime required

[Steve Lhomme <[e-mail]>]

To follow up on this, the specs should probably be clearer about this.

Since there are cases where the chapter end time is not needed, it
can't be mandatory. But in the ordered chapters case, it cannot work
if a chapter end is missing. Although the end may come from the
chapter parent if it has one...

2018-03-28 18:58 GMT+02:00 hubblec4 <[e-mail]>:
> Hi all
>
>
> Am 28.03.2018 um 17:51 schrieb Dave Rice:
>>
>> Hi cellar,
>>
>> This is to follow up on a discussion on ffmpeg-devel, see:
>> http://ffmpeg.org/pipermail/ffmpeg-devel/2018-March/227453.html. In the
>> Matroska Schema, the ChapterTimeEnd element is optional, but in
>> https://github.com/Matroska-Org/matroska-specification/blob/master/chapters.md#editionflagordered
>> it says "If an Edition of Ordered Chapters is enabled then the Matroska
>> Player MUST play those Chapters in their stored order from the timecode
>> marked in the ChapterTimeStart Element to the timecode marked in to
>> ChapterTimeEnd Element.”
>
> Thats right and for a "normal" ordered edition the endtime stamps are
> required. What means "normal": This is the case when ChapProcessCodecID is
> set to 0(default 0).
> If this value is set to 1 then Matroska-DVD-Menu system is used and there is
> it possible to have empty endtime stamps.
>
>>
>> So is ChapterTimeEnd required when EditionFlagOrdered=1?
>
> Yes.
>>
>>   Is the second edition of the example below (from
>> https://archive.org/download/chapters_test/chapters_test.mkv) invalid to
>> write or invalid to read?
>
> I'm sure Mosu's Chapter Editor(Matroska Reader) can read such mkv's but a
> Matroska Player will fail to build a correct virtual timeline.
> LAV Splitter ignores all ordered chapters with a play-duration of 0 or less.
> An empty endtime stamp is like a value of 0 so you get a negativ chapter
> play duration.
> Chapters_test.mkv  plays the first editon, but when you change to edition 2,
> playing is stoped, timeline greyed out, play-button is disabled and so on.
> There is no virtuell timeline for the second edition.
>
> Read and write is maybe possible and not really invalid, but for practice
> using is it invalid.
>
>> Should we state a fallback end time (segment duration?) when there is no
>> ChapterTimeEnd in an Ordered Edition?
>
> A fallback which used the segment duration is possible only when you have
> only one chapter. Otherwise the new virtuell playtime is too large and
> duplicates video content.
> The Segment duration is not a "good" value because this element is not
> always present AND (very)often too large.
>
>>
>> |+ Chapters
>> | + Edition entry
>> |  + Edition flag ordered: 1
>> |  + Edition flag hidden: 0
>> |  + Edition flag default: 1
>> |  + Edition UID: 3475548369
>> |  + Chapter atom
>> |   + Chapter time start: 00:00:15.000000000
>> |   + Chapter time end: 00:00:20.000000000
>> |   + Chapter flag hidden: 0
>> |   + Chapter flag enabled: 1
>> |   + Chapter UID: 12865469183194029579
>> |   + Chapter display
>> |    + Chapter string: Random Red Ball
>> |    + Chapter language: eng
>> | + Edition entry
>> |  + Edition flag ordered: 1
>> |  + Edition flag hidden: 0
>> |  + Edition flag default: 0
>> |  + Edition UID: 12338659363134957115
>> |  + Chapter atom
>> |   + Chapter time start: 00:00:00.000000000
>> |   + Chapter flag hidden: 0
>> |   + Chapter flag enabled: 1
>> |   + Chapter UID: 17174098126947929771
>> |   + Chapter display
>> |    + Chapter string: Full
>> |    + Chapter language: eng
>>
>> Dave Rice
>> _______________________________________________
>> Cellar mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/cellar
>
>
> Kind regards
> Martin Below
>
>
> _______________________________________________
> Cellar mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/cellar



-- 
Steve Lhomme
Matroska association Chairman