Hi All, I have been developing a project that uses Ogg Theora and Vorbis. As part of it I need to join various ogg streams to make a single stream. There are various use cases for this but lets use advert insertion as an example. I can vary tightly control all the properties of the Audio and Video streams, so all the input parameters in the setup headers of the Vorbis and Theora are the same e.g. Frame size, Frame Rate & Sample Rate. In my experiments I have been able to successfully join various streams by simply remultipexing the Theora and Vorbis streams in to a single ogg stream. For this to work I had to remove all the setup headers from all but the first stream. I also had to recalculate the Granule Pos time codes to make these match the new position in the stream. Also since some streams have the audio and video terminating at slightly different points they can cause the subsistent streams AV to become out of sync. So I had to repeat the last frame to make it fit the audio stream more length as close as possible. The results of the experiments were very promising. The files played in all the common Theora players (VLC, GStreamer & MPlayer) with no AV Sync issues. They also worked in players which determine the length of a Theora stream accurately (by decoding the last Granule Pos). This process is fast and efficent and has no impact on the quality of the output. My questions is how will Thusnelda effect this method of joining. If I all the input streams are encoded with the same version of Thusnelda, will the headers will be same or will they change. My understanding is there are structures in the headers which have been constant in those headers in all version of Theora which may well change in Thusnelda release. Will they be constant to one particular version of the encoder, or do they change on a stream by stream basis. Thanks, Charlie M