This directory contains "template" setups for some common benchmarking
tasks.

These depend on audio databases from the LDC
(http://www.ldc.upenn.edu/) which are quite large and cannot be
redistributed.  Where possible we have added notes on the exact LDC
catalog numbers and required files.

Some of the contents are inspired by Keith Vertanen's WSJ recipe
(http://www.inference.phy.cam.ac.uk/kv227/sphinx/)

-- dhuggins@cs.cmu.edu, 2007-06-25

