This is about to start another time around the same circle, but if the
arguments need to be restated, I'll take a turn on this lap.
We fully share the points 1) and 2) stated in the e-mail below from
Cullen since implementing and deploying a new codec in networks
(gateways, service plate-forms, mediaservers...) and in terminals
represents high costs for service providers, manufacturers and chipset
providers in terms of development, deployment and testing with risks
to create bugs and problems affecting customers. Furthermore, this
multiplies the problems of interoperability with already deployed
codecs and the transcoding needs to be addressed with related costs
(gateways) and quality degradations.
Although it is something of an aside, I find it difficult to believe
that this would ever be seriously considered within the ITU as a
reason why a clearly better codec should not be implemented and
offered for sale*[1].
I also find it somewhat difficult to believe, despite the real costs
of adding another row/column to full-matrix transcode, that commercial
providers would become very cross at the idea of a higher performance
standardized codec for free.
Therefore, the 3 stages mentionned are essential to be run sequentially:
"(1) get consensus on the requirements, (2) see if an existing codec
meets the requirements, and (3) specify a new codec only if none are
found in stage 2. Initially the WG would be chartered for (1) and when
that was done it would be re-charted for (2) and so on. "
Dibs on the first donut at the meeting planning meeting pre-meeting!
To address the points:
1) yes, we need a consensus on requirements. We can synthesize
requirements as a low-bar aggregation of the performance numbers of
already-demonstrated best of breed. At least, that's a decent enough
empirical strategy. We know what the current commercial codecs can
do, we know what the available free codecs can do that will be inputs
to this process and we know that there's demand for unencumbered
deployment. The requirements doc, as screenwriters might say, "writes
itself" (the union of what we can do and what the world needs). It
does need to be done of course, and screenwriters can't live without
thoughful editors, but what part of this is actually controversial?
2) we are here because the lack of a suitable standardized codec has
already been demonstrated. The driving impetus behind the proposal is
only partly technical. If, for example, the ITU wanted to view this
as an insurrection to be crushed*[2], it could do so by offering
suitable current-generation codecs without licensing strings, but we
believe this isn't possible and it's not even fair to offer as a
straw-man. Further, we believe that it is also not possible to start
fresh and arrive at a royalty-free result solely within the ITU or
MPEG or elsewhere as there are no examples of such. That said,
collaboration is highly desirable. I don't presume to posess
technical insight superior to the ITU.
However, you suggest:
Once the first requirement establishment stage completed, the working group
will then communicate detailed description
of the requirements and goals to other SDOs including the
ITU-T, 3GPP, and MPEG to jointly analyse and determine if existing
standardized codecs meet the requirements or can be efficiently adapted
to meet them. This work will constitute a second stage that will be
discussed and agreed with the relevant SDOs and the WG Charter will be
updated accordingly
...which effectivly amounts to first stopping dead, and then ceding
large chunks of the decision making process to outside groups with no
previously demonstrated interest in our goals. This doesn't make much
sense. 'Collaboration with the IETF' becomes a
joint-venture-by-committee at best. Though it is important to be in
active communication with other SDOs, the idea of proceeding within
the IETF is because IETF participants have expressed interest where
other SDOs haven't. I propose that we do not wait for any and all
passengers to meander to the station in their good time; those who
show up before the train leaves according to schedule are welcome to
get on board*[3]. There will be multiple stations along the route, of
course.
3) Several free codecs have been demonstrated that fufill basic
operating requirements. However, we have a unique opportunity to
leverage the strengths of several members of the field into an even
stronger contender. As stated multiple times in this process, we are
not interested in rubber stamping. I would consider it a wasted
opportunity. We have the rare spectacle of multiple groups of experts
willing to work together to produce a new, state-of the art audio
codec (or codecs) with no royalty or licensing strings attached at the
same time there is demand for exactly this! There is demand, there is
interest, there is motivation.
Monty
Xiph.Org
[1] one would expect the result to be a commercial product
[2] "You are a Rebel and a Traitor! Take her away!" I always thought
it was better in David Prowse's own voice honestly.
[3] Please mind the doors. Another train will doubtless be along presently.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf