[foms] Proposal: adaptive streaming using open codecs

[Steve Lhomme]

On Wed, Nov 17, 2010 at 7:15 PM, Timothy B. Terriberry
<tterriberry at mozilla.com> wrote:
>> Can you explain what's a LSP extrapolation ? A quick google doesn't
>> give anything.
>
> Probably because I meant LPC (linear prediction coefficients), not LSP
> (line spectral pairs). The concepts are related.
>
> Vorbis, for example, already uses LPC extrapolation to pad the first and
> last MDCT blocks, since blocks are overlapped, but there's no real data
> to overlap with at the beginning and end of a track. Code starts at
> https://trac.xiph.org/browser/trunk/vorbis/lib/block.c#L416

That code gives no indication of what vorbis_lpc_predict() does. Do
you really think that thing is easier than a cross fading ? Pretending
that a low quality stream is like the higher quality is outside the
scope of this discussion.

Anyway, I think the stream transition is player/browser dependent.
Everyone will have their own technique to make it smoother. All we
need at this level is to know what to expect from the various chunks
in terms of stream properties and timecodes.

And speaking of timecodes, Using a manifest of chuncked data makes it
very easy to insert commercials without changing anything in the
original stream. That means the timecodes of each chunck should not be
assumed to be continuous ? Same thing if you want to make a different
edit of the same content (different order, different ending, etc).

Steve