[NUT-devel] Re: [Vorbis-dev] minor Vorbis specification amendment proposal

[Rich Felker]

On Thu, Jul 20, 2006 at 08:58:41PM +0200, Michael Niedermayer wrote:
> Hi
> 
> On Thu, Jul 20, 2006 at 09:26:24AM -0700, Ralph Giles wrote:
> > On Thu, Jul 20, 2006 at 03:56:48PM +0200, Michael Niedermayer wrote:
> > 
> > > Vorbis has 3 "global" header packets, while generic containers generally
> > > have only a single global header per stream
> > 
> > I think this is a useful idea, but giving specific parse rules to try on 
> > *any* container format seems unwise. 
> 
> the suggestion was only for containers which support exactly one global
> header, for them i see no problem, maybe you could elaborate on what you
> think is unwise?

.....

> > I'd rather see an appendix 
> > documenting a standard embedding proceedure for various other 
> > containers, and perhaps some guidelines for creating new mappings 
> > including the parsability information you included.
> > 
> > Would you be willing to write up and contribute something like that?
> 
> sure, what else over what i already proposed would be needed?

imo you should make it clear that this amendment is in regard to how a
decoder implementation should parse a single-block header packet
received from a generic demuxer, not what every demuxer should do in
order to specially-accommodate vorbis. not sure if they're clear on
that matter or not.

rich