xiphmont@xiph.org
2007-Aug-08 04:43 UTC
[ogg-dev] Embedding int32's within threora & vorbis files
On 8/8/07, Ian Malone <ibmalone@gmail.com> wrote:> On 08/08/07, Paul Griffiths <gafferuk@gmail.com> wrote: > > Hi, i need to include a discontinuous stream of in32's(3 at a time) to vorbis > > and theora ogg files, to be played back on time. > > > > Does anyone have an example I can work from? > > > > I would love to say what it is for but I have to be careful of comptition at > > this early stage. > > > > If I understand correctly you want to store integers > in the Ogg stream at some synchronisation with > other contents (Vorbis, Theora).Correct. You define a discontinuous 'codec' and a 'mapping' that specifies how it is embedded in the Ogg stream. This makes sure it is only used by software that recognizes what it is.> Possibly OggPCM > would suit,No. OggPCM is, I think, defined to be continuous and definitely defined to be audio.> you probably want to look at defining an > application specific mapping so that no player would > attempt to play the (presumably) non audio stream.Exactly. Monty
Ian Malone
2007-Aug-08 05:01 UTC
[ogg-dev] Embedding int32's within threora & vorbis files
On 08/08/07, xiphmont@xiph.org <xiphmont@xiph.org> wrote:> On 8/8/07, Ian Malone <ibmalone@gmail.com> wrote: > > On 08/08/07, Paul Griffiths <gafferuk@gmail.com> wrote: > > > Hi, i need to include a discontinuous stream of in32's(3 at a time) to vorbis > > > and theora ogg files, to be played back on time. > > > > > > Does anyone have an example I can work from? > > > > > > I would love to say what it is for but I have to be careful of comptition at > > > this early stage. > > > > > > > If I understand correctly you want to store integers > > in the Ogg stream at some synchronisation with > > other contents (Vorbis, Theora). > > Correct. You define a discontinuous 'codec' and a 'mapping' that > specifies how it is embedded in the Ogg stream. This makes sure it is > only used by software that recognizes what it is. > > > Possibly OggPCM > > would suit, > > No. OggPCM is, I think, defined to be continuous and definitely > defined to be audio. >That makes sense. The spec for this new codec could probably be copied almost wholesale from OggPCM. The only obvious change needed is to redefine sample rate to simply providing a granulepos mapping. -- imalone