[vorbis] The immediate future

[Thomas Marshall Eubanks]

Angus Griffin wrote:

> I'm curious to know what people see in CBR. Not only in ogg, but in mp3's
> too, and especially in broadcast's such as icecast etc. Can someone
> enlighten me as to why one would want to use CBR when VBR is available,
> please?
>
> Angus
>

Hello;

   Because real pipes are bandwidth limited.

  Suppose that you want to unicast or multicast  UDP to 2-channel ISDN
listeners, who
are limited to 128 kbps. In MP3 CBR, you can use 90 kbps CBR and, if all goes
well, the listener will receive an uninterrupted stream, with about 38 kbps to
spare.

With VBR, even if the mean bit rate can be adjusted to be 90 kbps, there will
be FOR
SURE times when the actual rate is > 128 kbps, and packets will be dropped. If
you use TCP (which means that those dropped packets will be resent), you run
the real
risk of running out of buffer room in many applications (and TCP doesn't
scale).

   On a related topic, in CBR MP3's there  is the concept of a  "bit bucket,"
which is a time averaged VBR. In the MP3 CBR, each packet is the same bit size
and at a constant rate. Of course, in real music or speech the number of bits
required to meet a given audio standard will vary. Since a MP3 frame (AKA
Application Data Unit or ADU) represents 1152 samples, this means that there
will be times when the  frame is too big, some times too small. With the bit
bucket there is a pointer which refers to the location of the start of the
actual frame within the packet, so that bits can be borrowed from one frame and
given to another, by  making the frame boundaries not be at packet boundaries.

In the streaming context, THIS IS A VERY BAD IDEA !!!!

It means that if one packet is lost, more than one frames are lost (and,
therefore,
other data are useless even if received.

Ross Finlayson of Live.com has removed this from MP3, creating a variant
discussed at the recent IETF-48 in Pittsburgh, and described at the
http://www.ietf.org/internet-drafts/draft-ietf-avt-rtp-mp3-03.txt , where
the frame (or ADU) is the same as a sent packet.

(see http://www.ietf.org/html.charters/avt-charter.html if the draft has been
replaced
by a RFC, which should happen any day now).

I would STRONGLY recommend that CBR for Vorbis would mean
that One Packet = one audio frame (or ADU).  This would mean that at times the
actual data rate would be LESS than the CBR, but that's OK as long as the rate
is never FASTER than the CBR.

                                   Regards
                                   Marshall Eubanks

   T.M. Eubanks
   Multicast Technologies, Inc
   10301 Democracy Lane, Suite 410
   Fairfax, Virginia 22030
   Phone : 703-293-9624
   Fax     : 703-293-9609

   e-mail : tme at on-the-i.com     tme at multicasttech.com

 http://www.on-the-i.com         http://www.buzzwaves.com

--- >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.