[Matroska-devel] mkv standardization and documentation

Moritz Bunkus moritz at bunkus.org
Fri Apr 17 08:27:06 CEST 2015


Hey Dave,

> The project selected several audiovisual formats to focus on and
> MediaArea wrote a proposal focusing on three of their selections:
> FFV1, LPCM, and Matroska. As part of the proposal we noted that
> Matroska and FFV1 have not undergone a formal standardization process
> and wrote up these plans [1] to attempt to facilitate this via the
> IETF. Last week PREFORMA announced that MediaArea’s proposal was
> selected to continue to the prototyping phase [2].

Like I said on IRC: that's great to hear :)

> 	In combing through the MKV documentation during the design phase we
> 	have a few recommendations for the existing documentation and want
> 	to start sending pull requests. At the moment our intent is not to
> 	change the current version of Matroska but to clarify, refine, and
> 	correct any issue in the existing documentation in order to get it
> 	into better shape to submit as an Informational Standard to the
> 	IETF. We hope to work closely with the community at matroska-devel
> 	during this, please let us know if there is any objection or
> 	concern.

Sounds good to me.

> 	It made sense to us to start a documentation review with EBML and
> 	then work towards Matroska’s documentation. With EBML I see there
> 	are two main documentation sets: a draft RFC at
> 	http://matroska.org/technical/specs/rfc/index.html
> 	<http://matroska.org/technical/specs/rfc/index.html> and the EBML
> 	sourceforce page at http://ebml.sourceforge.net/specs/
> 	<http://ebml.sourceforge.net/specs/>. The two sets seem that have
> 	diverged at some point; for instance the RFC draft labels element id
> 	0x42F2 as EBMLMaxIDWidth whereas the sourceforge site calls is
> 	EBMLMaxIDLength. The definitions for the CRC Element vary between
> 	the two as well in respect to ordering. So does either the RFC draft
> 	or the specs at Sourceforce take precedent over the other? What is
> 	the status of the RFC draft of EBML?

The RFC draft is work someone else started way back when, but it never
got real attention, nor was it properly reviewed or even finished. The
RFC is definitely not authoritative; the specs should be. Although I
have to admit that I haven't looked at the specs EBML specs at
http://ebml.sourceforge.net/specs/ in ages, hence me saying that they
_should_ be authoritative.

What I am usually working with is the Matroska specs page at
http://www.matroska.org/technical/specs/index.html – which seems to
include all the information from the EBML specs page.

I think that working out the differences between the two specs pages and
the RFC and coming to a consensus how to deal with those differences
would be a good starting point. From there finishing the RFC would
likely come next.

I would also like to reduce the number of pages listing specs, most
likely resulting in removing the stuff from sourceforge. This would cut
down on the confusion and lead to fewer mistakes and discrepancies.

Kind regards,
mosu
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.matroska.org/pipermail/matroska-devel/attachments/20150417/022621b2/attachment.sig>


More information about the Matroska-devel mailing list