[vorbis-dev] Re: [theora-dev] Re: RFC3533

[Ralph Giles]

On Wednesday, May 21, 2003, at 10:35 am, Arc wrote:

> Correct me if I'm wrong, but I also recall Monty immediatly following
> this by suggesting this is a temporary restriction and is likely to be
> removed in the future.

That's looking less likely now that it's in an RFC. :)

Seek efficiency is the other main reason for this limitation. Since 
most codecs (that we write anyway) are dependent on the initial couple 
of packets to properly set up the decoder, you have to know where that 
is before you can seek to the middle of a chain segment. Having all 
those initial packets all at the beginning of the of the chain segment 
means you can find them with a bisection search instead of having to 
scan backwards for  the matching bos page.

How much of an efficiency win this might be depends entirely on the 
application of course. Your idea doesn't sound like it would need to 
search much at all, so it's all lose. When I first encountered Ogg I 
thought it was a pretty silly limitation too, but I've come to 
understand it as a useful limitation for Ogg's intended purpose as a 
multimedia stream format.

But as Arc suggests, you could use the ogg design without this 
restriction as long as you clearly document it as such. There's 
unlikely to be enough application overlap to generate confusion.

  -r

--- >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.