ietf
[Top] [All Lists]

Re: RTCWeb proposal

2013-12-04 08:42:18
On Wed, Dec 4, 2013 at 9:08 AM, Yoav Nir <synp71(_at_)live(_dot_)com> wrote:

I would first ask the potential implementers what they would do if two
endpoints negotiated down to motion JPEG, which according to Wikipedia,
means 320x240 at 12 FPS using up over 1 Mbps, which is more than the
upstream capacity in many places.

Would they show the user this barely-usable video stream, or would they
show a message that video couldn't be made to work?



I am on several fat pipes here and I get those messages all the time on
Skype etc. etc. So I don't see that this would be introducing a completely
new concern.

The endpoints are only going to select motion JPEG if they have no better
shared codec available. So this is only going to occur in a case where the
applications would otherwise be unable to communicate.


The only essential criteria I have for MTI is that it stands a chance of
being implemented universally which in this case means that it has to be
definitively unencumbered. Microsoft is not going to risk a billion dollar
damages award to let Mozilla get the codec of their choice. Nor are they
likely to buy a second license to the necessary IP. Mozilla is not going to
pay for a H.264 license. No decision made in the IETF is going to change
the position on either side.

So making AV8 MTI is not going to increase interoperability, nor is making
H.264 MTI. Whatever the intentions of the inventors, there are numerous
patent holders that might make a claim.


The other option that appears to meet the definitively unencumbered
criteria is H.261.


Yes, the lowest common denominator is going to be markedly less efficient
than the encumbered codecs. That is not an argument for creating a new
process that allows a group of people who don't bear the costs of the
decision to decide a winner.

-- 
Website: http://hallambaker.com/
<Prev in Thread] Current Thread [Next in Thread>