[theora] oggcat or other ways to combine theora videos (with or without re-encoding)?

[Gregory Maxwell]

On Sun, Sep 20, 2009 at 9:23 AM,  <yorn at gmx.net> wrote:
> Gregory,
>
>> It's not clear to me what you're referring to by "packet number".
>
> I mean the ogg packet number, that is defined within the ogg packet structure (ogg.h - http://www.xiph.org/ogg/doc/libogg/ogg_packet.html ):
[snip]
>
> These packets are returned by theora and vorbis encoders:
[snip]
>> Ogg has two relevant framing constructs, the page and the packet. Ogg
>> packets are not explicit numbered.
>
> Aren't they? What's that packetno-field stands for?

No.

> BTW: The _page_ numbers are correct, as far as I tested that, but this is done by libogg not by the encoder.
[snip]

The particular field is used only on the decode side to pass
information about holes in the stream back to the decoders from
libogg. This is done because libogg can detect holes in the stream
because it operates at the page level which the codecs are totally
oblivious to, but they would not be obvious at the packet level
without translating some of the per-page data into per-packet data.

On the encode side this packet number field is completely ignored: