[foms] MPEG TS streams and patents

Mark Watson watsonm at netflix.com
Tue Dec 14 12:35:08 PST 2010


HJ,

I think you confuse slightly MPEG and the MPEG Licensing Authority. The latter quite rightly attracts some ire due to its handling of some patent pools. MPEG develops both royalty-bearing and royalty-free standards. The mp4 file format, for example, has been quite successful and I think DASH is much closer to that work than to the codec work.

We have been at the same meetings and I have a very different impression. Yes, there have been a lot on unnecessary contributions but these have generally been rejected. Some research organizations come with proposals which do not have strong market requirements and these tended to be rejected as well. But this can happen anywhere. A lot of good topics have been discussed in detail and some good solutions agreed. I am not sure what you expect from a standardization process: there are always going to be some long discussions as people arrive with different objectives. That's not lack of goodwill, just reality. The point of the exercise is to harmonize these.

DASH published their Committee Draft in record time (4 months from CfP responses to CD) - I don't think this is "unconstructive". Also, noone can guarantee that a standard is royalty-free. The best you get is a commitment from those actually involved in the standards group.

I am not saying there are no issues that might lead a group like FOMS to follow a different approach, but I don't think the characterization below is accurate at all.

Regards,

Mark


On Dec 13, 2010, at 4:58 PM, 이현재 wrote:

> I have same concern with Jereon.
> As we all know, MPEG has been very hostile mud, struggling to get whatever
> tiny IPR even though unnecessary. Paradoxically speaking, due to the tragic
> success of MPEG2 adoption of broadcast system, it's natural gene is
> corrupted.
> DASH discussion is very unconstructive as far as I can tell from attending
> the meeting so far.
> 
> I generally agree with Mark in that it may be royalty free because we have
> a few powerful free alternatives, but still I have lingering concern
> because CE industry has bad experience of MPEG based mp3 and jpeg royalty
> claim that were regarded royalty free when other alternatives coexisted.
> Basically CE industry has a dormant phobia to MPEG. It does not guarantee
> royalty free in written document.
> However, MPEG is weakening now, besides MPEG2, MPEG4 is going to royalty
> free because of powerful free alternatives.
> 
> I would like to mention that the progress of DASH is unsuccessful so far
> because of its bad gene. Somebody just say many people and many
> contributions are the metric of successful discussion which I think not. If
> we start the discussion at different place, it will enjoy good progress
> based on good will.
> 
> Best regards,
> HJ 
> 
> 
> 
> Message: 3
> Date: Mon, 13 Dec 2010 09:38:55 -0800
> From: Mark Watson <watsonm at netflix.com>
> Subject: Re: [foms] MPEG TS streams and patents
> To: Foundations of Open Media Software <foms at lists.annodex.net>
> Message-ID: <7781D34C-C357-4FFB-B9AD-DF5BDCD96345 at netflix.com>
> Content-Type: text/plain; charset="us-ascii"
> 
> All,
> 
> I would not recommend using Transport Streams - and not just because of the
> possible IPR.
> 
> MPEG DASH supports both MPEG2 TS and MP4 file formats. It doesn't mandate
> either and you could easily specify how the DASH manifest could be used
> with a different file format such as WebM, provided the file format has the
> necessary features. There is also nothing mandated in DASH which is
> specific to a particular media codec, though obviously MPEG tries hardest
> to provide support for MPEG codecs.
> 
> In any adaptive streaming standard, manifest format, media file format and
> media codecs should be decoupled as far as possible.
> 
> Most important for adaptive streaming is agreement on the file format and
> general data model. Specific manifest format is secondary, because if you
> have an agreed data model it can easily be mapped into your favorite format.
> 
> IMO, a good way for someone to completely kill MPEG DASH would be to
> attempt to charge royalties - there are "free" alternatives. Everyone I
> have spoken to on the committee understands this and I have not heard any
> talk about folding DASH into any patent pools (existing or new).
> 
> I can't say for certain that noone will, but it makes no sense for people
> to invest in developing a standard and then to take action which they know
> will kill it. I think it very likely it will end up with the same status as
> ISO Base Media File format, which is widely used without IPR issues.
> 
> Obviously, though, you have to make your own judgements: "Mark thinks it is
> very likely..." means exactly what it says.
> 
> ...Mark
> 
> 
> 
> On Dec 13, 2010, at 5:21 AM, Jeroen Wijering wrote:
> 
>> Hello all,
>> 
>> Just to be sure, a transport stream is also a patent encumbered format,
> correct? So packing up VP8 / Vorbis into TS fragments for adaptive
> streaming would not be an open solution? It looks like transport streams
> are part of either the MPEG2 or MPEG2 Systems pool from MPEG-LA. See the
> PTS/DTS/PAT/PMT references: 
>> 
>> http://www.mpegla.com/main/programs/m2/Documents/m2CrossRefChart.pdf
>> http://www.mpegla.com/main/programs/m2s/Documents/m2sCrossRefChart.pdf 
>> 
>> Thinking forward, would this mean MPEG DASH is also a no-go as an
> adaptive streaming standard? It seems to rely heavily on TS / MP4 / AVC. In
> turn, DASH might also be folded into one of the MPEG LA pools. Does anybody
> know more about this?
>> 
>> Kind regards,
>> 
>> Jeroen
>> _______________________________________________
>> foms mailing list
>> foms at lists.annodex.net
>> http://lists.annodex.net/cgi-bin/mailman/listinfo/foms
>> 
> 
> 
> 
> ------------------------------
> 
> _______________________________________________
> foms mailing list
> foms at lists.annodex.net
> http://lists.annodex.net/cgi-bin/mailman/listinfo/foms
> 
> 
> 
> _______________________________________________
> foms mailing list
> foms at lists.annodex.net
> http://lists.annodex.net/cgi-bin/mailman/listinfo/foms
> 



More information about the foms mailing list