[foms] WebM Manifest

[Thomas Vander Stichele]

On Thu, 2011-05-05 at 19:22 +0200, Steve Lhomme wrote:
> I understand everyone would like a single solution that works in all
> cases. But what would you prefer ? 2 systems that are optimum for a
> single case or 1 system that works OK but no great in all cases.
> 
> t's not like servers or players don't already have the "choice"
> (rather necessity) to support more than 1 adaptive streaming format.
> So I don't think it's a big deal if there is one for VOD and one for
> live.

I'm not sure I understand this point.

- http flv streaming works the same for ondemand and for live from the
player pov
- same applies to mp3
- same applies to ogg
- same applies to asf
- same applies to webm single bit rate
- same applies to HLS, smoothstreaming, F4V
- same applies to RTSP, Real, mms, ...

Why would you want WebM, a minority codec that people already don't seem
to want to deploy anything for, to have two different systems ?

It's taking long enough for Firefox to support Ogg live well enough, or
Chrome to support live WebM well enough, when the changes to make it
work were relatively minor.

Why would we want to support two alternative systems for live and
ondemand ?

There are people working on a blend of live and ondemand as well; in
essence, should a time-indexed 'live resource' (what the server was
streaming 30 minutes ago) be using 'live' or 'ondemand' technology ? Why
can't it be the same ?

>  After all these uses cases would never recoup anyway. A system
> optimize for live could for VOD for sure. But a VOD system doesn't
> need to be hacked and made complex for the sake of somehow working for
> live streaming.

That's a strawman's - obviously it's not something that's more complex
or hacked.  It's definitely less complex than supporting two separate
implementations.

Thomas
-- 

-- 
What part of punching you in the face did you not understand ?
--
savon - Saving your work to svn
https://apestaart.org/thomas/trac/