Getting a First Feature Description File

If you want JANUS (or the featmaker program) to read speech recording files you have to tell it how to do that. Usually the speech files are a sequence of plain 8 or 16 bit samples, possibly preceeded by some header. Before you can do anything with the speech files you'll have to write (or take from somewhere) a feature description file, where you describe what kind of data can be found in the speech files and how to preprocess them before they are used. If you've never seen a feature description file before, you might be served best by simply taking an existing one from a library of feature description files, or by asking somebody who is familiar with them.