Are speech technologies mature enough?
Speech synthesis and recognition technologies have achieved significant progress over the last years. The annual world market of speech technologies exceeds 1bn$. Speech synthesis is part of our day-to-day environment with many different embedded systems, including domestic appliances. Text-To-speech (TTS) software have dramatically improved in terms of voice quality, continuous speech flow, and the support of many configurable parameters (voice type, pronunciation dictionaries and rules, speed, intonation, etc). It then sometimes becomes hard to differentiate a synthesized voice from a natural one. Recent speech recognition capabilities improvements are even more dramatic. Automatic Speech Recognition (ASR) algorithms are now able to immediately cope with any caller without the need of any learning phase. ASR can now understand natural language and recognize proper names in huge directories with high levels of recognition rate. In addition, the new release of the VXML standard creates
