[foms] Text of ISO/IEC DIS 23001-6 Dynamic Adaptive Streaming over HTTP

Jeroen Wijering jeroen at longtailvideo.com
Tue Feb 15 07:39:24 PST 2011


Hello all,

I wanted to forward a couple of updates on the WebTV workshop. They've been two interesting days, especially on the adaptive streaming / content security front.  We had a session on both adaptive streaming (focussing on DASH) and on content protection (focussing on PIFF). Sheets of the presentations done at both sessions can be downloaded here:

http://www.w3.org/2010/11/web-and-tv/agenda.html


On the DASH side, the spec has just been released. Silvia has forwarded a download link already (see below). In short, it's a flexible XML format that seems to support all required use cases (fragments, range-requests, manifest polling, etc). Outstanding issues are:

*) Licensing. What happens here is not entirely clear. At their presentation, Microsoft committed to make available any patents in this area they have under the MPEG RAND license. The W3C is still going after other contributors to the spec, to see if they want to take similar steps.
*) Profiling. Since the system is quite flexible, it is possible to define profiles for DASH. Profiles can contain both limitations to the XML elements / attributes and the format of media files referenced in the XML. A "basic ondemand" profile is defined; there'l probably be a "basic-live" profile as well. 

I attached example DASH files that use WebM to define both an ondemand and live stream. The ondemand one uses range requests to pull both the Cues and fragments. The live one uses reloading of the manifest (like Apple does), a URL template (so a full fragment list is not needed) and separate codec setup segments (so they're not constantly duped - I understand setup in Vorbis is quite big). They should give a rough idea of how an ondemand or live stream can be defined with a DASH manifest. 


On the content protection side, there was some talking about PIFF - Microsoft's encrypted, fragmented MP4 format. Interesting about this is the clear separation of encryption/decryption and the rights management layer. All presenters backed this approach. Perhaps this is interesting for open video / WebM as well, since encryption can be very useful for e.g. basic content protection (no DRM, just keys) or privacy reasons.  Any real standardization seems a bit further away though. There's been no detailed proposal on how this decryption / key signalling could work in HTML5 (cross-browser and cross-codec).


Steve Lhomme, JB Kempf and Mark Watson were also there; please chime in if I missed anything interesting. 


Kind regards,

Jeroen


-------------- next part --------------
A non-text attachment was scrubbed...
Name: basic-webm-live.xml
Type: application/xml
Size: 1804 bytes
Desc: not available
Url : http://lists.annodex.net/cgi-bin/mailman/private/foms/attachments/20110215/77102cb6/attachment.xml 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: basic-webm-ondemand.xml
Type: application/xml
Size: 1107 bytes
Desc: not available
Url : http://lists.annodex.net/cgi-bin/mailman/private/foms/attachments/20110215/77102cb6/attachment-0001.xml 
-------------- next part --------------


On Feb 13, 2011, at 1:02 PM, Silvia Pfeiffer wrote:

> FYI: MPEG have released their draft DASH standard for adaptive HTTP streaming.
> Plus there is a group forming in the W3C about Web & TV, which is also
> interested in the topic and looking at DASH with lots of interest.
> Cheers,
> Silvia.
> 
> 
> ---------- Forwarded message ----------
> From: Thomas Stockhammer <stockhammer at nomor.de>
> Date: Sat, Feb 12, 2011 at 6:14 AM
> Subject: [W3C Web & TV] Text of ISO/IEC DIS 23001-6 Dynamic Adaptive
> Streaming over HTTP
> To: public-web-and-tv at w3.org
> 
> 
> Experts,
> it was pleasure to meet you all.
> 
> I wanted to make you aware that the
> Text of ISO/IEC DIS 23001-6 Dynamic Adaptive Streaming over HTTP
> is now available here
> http://mpeg.chiariglione.org/working_documents/mpeg-b/dash/dash-dis..zip
> Note that this is a draft standard.
> Best regards
> Thomas



More information about the foms mailing list