702122 - add support for DASH (WebM)

Status: RESOLVED WONTFIX

(Core :: Networking, enhancement)

Description

[Josh Aas]

We should consider adding support for DASH (WebM) as an adaptive streaming solution.

MPEG DASH specification (in progress):

http://mpeg.chiariglione.org/working_documents/mpeg-b/dash/dash-dis.zip

DASH for WebM:

http://sourceforge.net/apps/trac/matroska/wiki/DASH_Profile

Comment 1

[Timothy B. Terriberry (:derf)]

Has the IPR situation around this spec become any clearer?

Comment 2

[Josh Aas]

(In reply to Timothy B. Terriberry (:derf) from comment #1)
> Has the IPR situation around this spec become any clearer?

Are you referring to specific concerns? If so, can you provide more information about them?

Comment 3

[Timothy B. Terriberry (:derf)]

When this was discussed at OVC for the past two years, one of the issues was that at least one vendor (Microsoft) was expected to have IPR covering the result, and possibly others, since there are IPR claims against other, competing manifest formats (see, e.g., https://datatracker.ietf.org/ipr/search/?option=document_search&document_search=draft-pantos-http-live-streaming for the list of IPR declarations against Apple's HTTP Live Streaming). I got assurances that the DASH WG members understood the result needed to be royalty-free to have a chance of competing against the multiple proprietary offerings in this space, but I haven't seen any official statement from Microsoft, or anyone else for that matter, that the result actually IS royalty-free.

Since claims like Apple's simply list "an unpublished pending patent application", it is impossible for us to verify what is actually claimed to see if DASH manages to steer clear of it. Apple still has the ability to modify the claims of their application before a final patent is issued.

Comment 4

[Ralph Giles (:rillian)]

In addition to Apple's claim against the Pantos draft, in which Emblaze state they intend license US Patent 6389473 under "commercial terms". At least that's a specific claim, so it's possible to evaluate whether whether is read on DASH.

Comment 5

[Mark Watson]

The link above for the specification may not be the latest (final draft) version. The latest version is available here: http://www.3gpp.org/ftp/tsg_sa/WG4_CODEC/TSGS4_66/Docs/S4-110852.zip

This archive includes the DASH specification itself as well as an amendment to the ISO Base Media File Format and anther document specifying common encryption for the ISO Base Media File Format.

The reason for the 3GPP link is that ISO have not published this draft directly, but have sent it to 3GPP, all of whose documents are public.

Comment 6

[Mark Watson]

Josh asked me to add my response to his question on DASH IPR to this thread:

My hope, based on conversations with the participants, is that it will be royalty-free. Indeed, if it is not, then it is probably dead on arrival - and people know this. But my hopes alone do not really answer the question, so here are some more details: 


Firstly, from an engineer's perspective (rather than a lawyer's) it is difficult to see how there could be much novelty in DASH specifically, given all the things which have gone before (Move Networks, Smooth Streaming, HLS, various proprietary deployments, including Netflix). One could expect companies to come forward saying they have IPR on some basic concepts of adaptive HTTP streaming that apply to DASH. But this doesn't help you choose DASH vs not-DASH for HTTP streaming because those same claims would apply to other systems also. For example Emblaze claim they have patents on Apple HLS (as do Apple) as noted above. 


Secondly, the ISO process is that each participating company with relevant patents must declare to ISO whether they will license those and on what terms. So far, Cisco [1], Microsoft [2] and Qualcomm [3] have made public statements that they will not seek royalties for DASH (Qualcomm's statement has a nuance that says something along the lines that they reserve the right to use DASH patents against people who try to use their core wireless technologies without a license - you should definitely read their statement rather than trusting my paraphrase.) The official statements to ISO are available at www.iso.org/patents . To my knowledge, noone has yet made a statement saying that they want royalties. 

For Netflix's part we have also stated publicly that we will not seek royalties for any IPR we may have on DASH. 

There is no process whereby the MPEG committee makes a definitive statement that something is Royalty-Free - you have to make your own judgement about the risk based on who is saying what. 

Finally, it's not unlikely that someone will come forward with something on one of the more obscure features in DASH - for example a side-player in the standardization who managed to get some feature included. A good solution there is just not to use that feature (just as noone uses any of the obscure features in the mp4 file format that have IPR claims). 

[1] http://blogs.cisco.com/sp/cisco-leads-the-way-in-standardizing-ip-video-delivery-through-mpeg-dash-specification/ 

[2] http://www.w3.org/2010/11/web-and-tv/slides/microsoft-john-simmons.pdf (Slide 4 - this is not a great reference, but they have made the commitment multiple times and Smooth Streaming is covered by their "community promise" - we can expect something more definitive from them soon). 

[3] http://www.qualcomm.com/blog/2011/08/16/dash-toward-better-mobile-video-user-experience

Comment 7

[Timothy B. Terriberry (:derf)]

(In reply to Mark Watson from comment #6)
> terms. So far, Cisco [1], Microsoft [2] and Qualcomm [3] have made public
> statements that they will not seek royalties for DASH (Qualcomm's statement
> [snip]
> For Netflix's part we have also stated publicly that we will not seek
> royalties for any IPR we may have on DASH. 

Thanks. This is exactly what I was looking for in terms of "the IPR situation around this spec becoming clearer".

> There is no process whereby the MPEG committee makes a definitive statement
> that something is Royalty-Free - you have to make your own judgement about
> the risk based on who is saying what. 

Yeah, the patent system makes such a statement impossible, so it's unreasonable to expect one.

Comment 8

[Max Kanat-Alexander]

  Any status updates on this (or an idea of what Gecko/Firefox release it might make)?

Comment 9

[Steve Workman [:sworkman] (INACTIVE)]

Added a new bug 734546 to track an initial code drop. First work in progress patches have been uploaded there.  Feedback would be greatly appreciated :)

Comment 10

[Steve Workman [:sworkman] (INACTIVE)]

DASH-WebM work in Firefox has been halted indefinitely due to increased industry momentum behind MediaSource plus JS-based implementations. Current functionality will not be enabled due to a desire to have one single approach to DASH support. For reference, support for MediaSource Extensions in Firefox is covered in Bug 778617.

Marking this as wontfix.