Re: [rtcweb] The MTI Codec Questions (what to ask and how to ask them)

[Alexandre GOUAILLARD <[e-mail]>]

Apple had 264, but there was no API to make 264 HW acceleration available
to developers. It was mentioned as one of the reason why cisco open264 felt
short of addressing some concerns voiced in the MTI pool. the iOS 8's 264
API changed this.

In any case, I support the chair decision to try to revisit the question.

On Tue, Nov 4, 2014 at 11:15 AM, Matthew Kaufman <[e-mail]> wrote:

>  A "substantive change" would be *any* of the parties who participate in
> these discussions having a different position than before.
>
> Yes, Cisco, who already supported H.264 and had announced an intention to
> ship an open source and binary module, shipped H.264.
>
> And Firefox, reluctant supporter of H.264 when available but preferring
> VP8 and some future even more free codec that isn't yet shipped, is
> reluctantly supporting H.264 when available.
>
> And iOS 8 has H.264, but of course Apple already had access to that H.264
> if they were to put WEBRTC into their browser...
>
> Etc.
>
> Where's the "Google has decided to pull their support for VP8 and fully
> back H.264 for real-time communications on the web" announcement? Or the
> "Microsoft open-sources Internet Explorer, adds native VP8 to Windows"
> announcement? Or the "MPEG-LA drops all fees for H.264 encoding and
> decoding" announcement? Or really *anything* that substantially changes
> things from where we were six months ago?
>
> I don't get the desire to spend time on this topic... and I especially
> don't understand the call to pack as many people into a room as possible to
> conduct a hum, when the outcome will need to be discussed and confirmed on
> the list anyway.
>
> Matthew Kaufman
>
>
> On 11/3/14, 5:13 PM, Jonathan Rosenberg wrote:
>
> Matthew,
>
>  You wrote:
> "I'm going to ask what I asked when I first saw this on the agenda: What
> has substantively changed since then?
>
> As far as I can tell, nothing.:
>
>  Actually several things have substantively changed since then. The list
> includes but is not limited to:
>
>  * Cisco shipped its H264 open source and binary module
> * Firefox is using the Cisco binary module and as such Firefox now
> supports both H264 and VP8
> * IOS8 has shipped with API support for H264 (you may recall that lack of
> a solution for H264 on IOS was an objection many had regarding the Cisco
> h264 binary module solution; this is now addressed)
> * there has been progress in the ongoing legal proceedings around VP8
> * there are new IPR statements regarding VP8 in ongoing standards processes
>
>
>  I think this is far beyond "nothing". And given the importance of this
> topic to progress on adoption of webRTC, it warrants discussion at the mic.
>
>  -Jonathan R.
>
>
> On Mon, Nov 3, 2014 at 8:02 PM, Harald Alvestrand <[e-mail]>
> wrote:
>
>>  Hm.
>>
>> I don’t think there’s much value in revisiting the “one codec”
>> alternatives. We tried that, and know that we found no consensus. Nobody’s
>> changed their minds.
>>
>> It would be very sad if we give up on interoperability for WebRTC
>> devices. Accepting “either” means that there will be 2 groups of them, and
>> they need a gateway to talk to each other, even when they can all talk to
>> all compatible browsers. Having an MTI would be better - but one purpose of
>> the “device” category is to allow fully compliant devices that don’t need
>> the kind of corporate backing a browser needs - which means that licensed
>> codecs are an issue. We’ve had that discussion before.
>>
>> Of course, WebRTC-compatible devices (as currently defined) can do
>> whatever they want.
>>
>> But still, it seems that there’s a chance that discussing this again is
>> worth it. We might find an agreement this time.
>>
>> Harald
>>
>>
>>
>>
>> On 11/03/2014 03:32 PM, Sean Turner wrote:
>>
>> All,
>>
>> One of the remaining major technical decisions for the RTCweb WG is which codec(s) should be  MTI.  The issue has been on hold for over 6 months and the original plan to was the re-attempt determining consensus at the IETF 91.  To make the best use of the WG’s face-to-face time at IETF 91, we want to give the WG ample time to digest/discuss the questions the chairs intend to ask the WG concerning the MTI codec (or codecs).  We want to know before the meeting whether to ask the questions and then what questions to ask - in other words we want to inform the WG of the questions before the WG session so as to not waste time debating what questions should be asked.
>>
>> Without further ado, these are the proposed questions:
>>
>> Question #0 (hum)
>>
>> Do you want to discuss this issue at this meeting?
>>
>> Question #1 (stand up)
>>
>> Please stand (or signal in the jabber chat) if you will be part of that consensus process for this question. If you're here to read email or watch the show, we want to know that your sitting throughout this isn't expressing opinions for the consensus process.
>>
>>     To many this might seem like a silly question,
>>     but the chairs believe the problem is well enough
>>     understood by those actively involved WG
>>     participants so we would like to confirm this
>>     understanding.  The chairs will also use to the
>>     determine the informed pool of WG participants.
>>
>> Question #2 (hum)
>>
>> Do you believe we need an MTI codec to avoid negotiation failures?
>>
>>     Previous attempts at determining the MTI did not
>>     yield a result but did confirm that there is a desire
>>     for an MTI to avoid negotiation failures.   Recently,
>>     some on the mailing list have expressed an interest
>>     in postponing this discussion until after IETF 91.  The
>>     purpose of this question is to reconfirm the original
>>     consensus.
>>
>> Question #3 (open mic)
>>
>> Are there any codecs that were not included in the previous consensus calls that warrant consideration?  If yes, which one and why.
>>
>>     The assumption is that the viable codecs are a) VP8,
>>     b) H.264, or c) VP8 and H.264.  This is based on the
>>     extensive poll results from the last consensus calls.
>>     But time has passed so we need to entertain the ever
>>     so slight possibility that another codec has miraculously
>>     appeared.  Remember, we want to ensure we’re going
>>     to get maximum interoperability.
>>
>> Question #4 (open mic)
>>
>> Are there any new or unaddressed technical issues that will not allow us to narrow the field to VP8 and H.264?
>>
>>     We do not want to revisit previous discussions; we only
>>     want new or unaddressed technical issues and will throttle
>>     the discussion accordingly.  We’ll rely on WG participants
>>     and our former RAI AD (Mr. Sparks) for help in this area.
>>
>>     We believe the technical discussion will fall into two
>>     buckets:
>>       - New or unresolved technical points.
>>       - Licensing.  WRT licensing, the IETF tries not discuss
>>         whether IPR is valid, but an IPR issue that can be used
>>         as input to the decision making process is if enough
>>         people say they can’t/won’t implement because of the IPR.
>>
>> Question #5 (hum)
>>
>> With respect to the MTI codec:
>>     - Who can live with a requirement that WebRTC User Agents
>>       MUST support  both VP8 and H.264 and WebRTC devices
>>       MUST support  either VP8 or H.264?
>>     - Who can live with a requirement that all endpoints MUST support VP8?
>>     - Who can live with a requirement that all endpoints MUST support H.264?
>>
>> Thanks for your time,
>> t/c/s
>> _______________________________________________
>> rtcweb mailing [e-mail]://www.ietf.org/mailman/listinfo/rtcweb
>>
>>
>>
>>   --
>> Surveillance is pervasive. Go Dark.
>>
>>
>> _______________________________________________
>> rtcweb mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/rtcweb
>>
>>
>
>
>  --
>  Jonathan Rosenberg, Ph.D.
> [e-mail]
> http://www.jdrosen.net
>
>
> _______________________________________________
> rtcweb mailing [e-mail]://www.ietf.org/mailman/listinfo/rtcweb
>
>
>
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb
>
>


-- 
Alex. Gouaillard, PhD, PhD, MBA
------------------------------------------------------------------------------------
CTO - Temasys Communications, S'pore / Mountain View
President - CoSMo Software, Cambridge, MA
------------------------------------------------------------------------------------
sg.linkedin.com/agouaillard

   -