The following depicts the directory structure and default partitioning proposed for the subcorpora on different discs. Subcorpora categories are denoted by the directory names in level 2. Different subcorpora will reside on different discs unless specified otherwise below. Training and development test data will probably be distributed together on one series of discs.
The different files for the 5k and 20+k test speakers can be distinguished between by means of the session number in the filename (see also section on File Naming Formats): 01 for all adaptation; 02 for the main session for training speakers; 02 for 5k for test speakers; 03 for 20+k for test speakers)