[Vorbis-dev] y0h

[Stephen Whiters-Ridley]

Thanks for your response. Ok, I will see how things progress, I am
tempted to fix the extensions, in which case I will submit them for
consideration. I looked briefly at py-ogg2. 'brief' because I was a
bit annoyed that there wasnt an official release of libogg2 just yet,
and that I had difficulty with the autoconf in my libogg2 checkout
from xiph.org....I will look at it more closely. Thanks again for your
response.

On 5/22/06, Ralph Giles <giles at xiph.org> wrote:
> On Mon, May 22, 2006 at 11:35:44PM -0400, Stephen Whiters-Ridley wrote:
>
> >                                                          ... essentially I
> > would like to instantiate a OggStreamState object, and use one of its
> > methods to push pages into the stream (encoding) to be ultimately
> > written to file. However, all the methods that look like they would be
> > responsible for creating packets/pages from raw data report that they
> > expect an object of type OggPacket, which is a class that doesnt
> > seemed to be definite
> > anywhere in pyogg. From looking at some of the sample "encoders" out
> > on the web, it looks like OggPacket creation is done actually done in
> > ogg.vorbis? is this correct?
>
> I've not looked in a long time, but iirc the python bindings just have
> huge chunks missing. They were written when vorbis was the only thing
> anyone was using in the ogg container and didn't properly separate the
> two layers.
>
> So, we'd love to have someone fix them up and/or rewrite the things to
> use a proper pythonic object model. (Again, IIRC) the current bindings
> just wrap the libogg api, which isn't really what you want in an OO
> language.
>
> BTW if you don't need pure python you might take a look at
> http://svn.xiph.org/trunk/py-ogg2/ and see if that's any better.
>
> Cheers,
>  -r
>