cicero 0.7.2-3 (amd64 binary) in ubuntu wily

 This Text-To-Speech (TTS) engine speaks French; a preliminary English support
 is also offered.
 The engine uses context-sensitive rules to produce phonemes from the text. It
 relies on MBROLA to generate actual audio output from the phonemes. The TTS
 engine is implemented using the Python programming language.
 .
 The upstream authors have come up with this TTS to try and meet their own needs
 as blind users.
 It's designed to be plugged as output to some screen-review software, firstly
 with BRLTTY.
 They favor speed and intelligibility over perfect pronunciation.
 Cicero is aimed to have a quick response time, the ability to quickly shut-up
 and skip to another utterance, intelligibility where it counts (not perfect
 pronunciation), the ability to track speech progression, relative simplicity
 (hackability) and relative small code size.