[vorbis] Bitrate stripping?
John Morton
jwm at eslnz.co.nz
Thu Jan 1 15:25:40 PST 2004
On Friday 02 January 2004 01:24, Daniel Schregenberger wrote:
> On Don, 2004-01-01 at 23:08, John Morton wrote:
> > On Thursday 01 January 2004 20:21, Graham Mitchell wrote:
> > > Once peeling is working out of the box, people will start using it
> > > (and loving it, I suspect), but for now there's not enough clamor for
> > > it to move to the top of the TODO list of anyone capable.
> >
> > And the longer it's left, the less useful it becomes, unfortunately.
>
> Why that?
Bit peeling is all about helping you out if you have limited storage,
bandwidth and cycles.
In the storage case, it's a win for owners of portable players, as it would
let you peel down your (say) q5 streams to q2 on the fly, so you can cram
more into the portable player - no transcoding, and no long term storage of
duplicate streams.
For streaming, it lets the stream source store one high quality stream and
allows them to present it as lower quality streams that are streamable in
realtime at 56k, 128k etc, and the bit peeling is, in principle, cheaper than
encoding three or four separate outputs for live material.
The problem is that the amount of storage, bandwidth and cycles you can buy
for a dollar is always increasing, and shows no sign of letting up in the
near future. Already the portables you can buy that that support vorbis come
in 20gb configurations, which is about the size of my entire album
collection, and I'm probably only a $1k computer upgrade away from making
storage in FLAC and transcoding to vorbis a viable proposition.
On the server side, the hardware with cycles enough to encode three or four
bitrate streams on the fly isn't all that expensive, and again storage isn't
too great a problem. In the mid term - say 5 years from now - I expect the
market for 56k streams to seriously contract, such that you'd really only
need to have 64k and maybe 96k streams for 128k bandwidth people, and q5-q7
streams for everyone else. That doesn't help your bandwidth problem at the
server end, but I expect that IceShare will be the killer app to solve that
problem.
Basically, two years ago bit peeling would have been a killer app -
potentially enough of a feature to override the whole 'free does not
compute' problem that vorbis has with the big corporate customers. Right now
it would be one of a number of nice features vorbis has, along with comments
that don't suck, easy replaygain support, better quality to size than most of
the non-AAC competition. It might even be that way in a year, but two years
from now? I think encoder optimizations, monster portables and IceShare will
make it irrelevant for most people.
John
--- >8 ----
List archives: http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body. No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.
More information about the Vorbis
mailing list