[vorbis-dev] q re vorbis packetizing in ogg stream

[Michael Smith]

At 02:07 AM 4/25/01 -0500, you wrote:
>>From looking at the code, it looks like the first vorbis header in a
stream is
>always placed in its own page.  It also looks like maybe the audio data
(i.e.,
>the stuff after the headers) is maybe also supposed to begin on a new page.
>Is this part of the spec?

Correct. The first packet must be on a page by itself (this is true for all
ogg streams, and is enforced by libogg).

The requirement that the audio data begin on a new page was added later, it
wasn't there originally. It's needed because otherwise streaming is overly
complex. Playback works regardless. This has to be done by the encoder,
because this requirement is specific to vorbis, not to ogg in general. I
believe this change was made prior to beta 3.

>I notice, though, that the second and third headers share a page, at least in
>one .ogg I looked at.  It seems like it would be easier for tools which
>manipulate the comment or codebook headers if each of these headers was
>required to be placed on its own page.  Any thoughts on this?

As someone who has written many (most?) of the available tools to
manipulate vorbis files in such ways, I feel qualified to answer this. It'd
make it _very_ slightly easier, but not significantly so - so it's not
worth changing the spec for this. It only makes the simple case slightly
simpler - and since one has to implement the complex case anyway (where the
packet doesn't fit onto one page - common for the codebooks, less common
but very possible for the comments), it's not a great help.

Michael

--- >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-dev-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.