The file format defines in which formal framework the specified data are embedded. Since a speech corpus always contains signals, symbolic data (annotations), meta data and - in most cases - a dictionary, we will describe those separately.