Here is a fictional use case for a feature I would like to see: Bob records a screencast to create a tutorial for PiTiVi. However, he wants to make it as accessible as possible by having it dubbed into multiple languages and released using an open format.

To do this he would need to:
- import a theora+vorbis video file
- record/import audio tracks for various languages
- render a new file that contains all the languages

However, there are a few conditions:
- the video stream must not be recompressed (unless asked by Bob) to keep the original quality
- the audio streams would all be saved in the final render, meaning that the end-user (viewer) will be able to change languages on-the-fly using his/her favorite media player software. For example, someone using Totem could go in the Audio>Language menu and select "Français" or "中文" or "English".

This would be a terrific feature.