“Embedded” speech technologies

In order to provide the most natural and efficient human-machine dialogue interaction, Tempest speech platform integrates the most advanced speech recognition and synthesis technologies.

The goal of Waycom, which has been reached with the development of the Tempest platform, was to make easy the use of technologies coming from various vendors and put simple but efficient tools in the hands of final users and system integrators.

The integrated technologies are producer independent (multi-vendor) and multilingual, for both the speech recognition (ASR Automatic Speech Recognition) and the speech synthesis (Text-To-Speech).

The platform is based on a “multi-tier” architecture, moreover it allows the distribution of speech resources on the local network or in the cloud, which offers the benefits of a high scalability in terms of contemporary voice channels, together with strong reliability and availability.

The typical complexity of speech interfaces and especially the one related to the speech recognition in telephony applications, is made simple and easily enabled through graphical interfaces and “wizard” that lead the developer in the design of interaction flows between the telephony platform and the final user.

In order to provide powerful and flexible tools to the developers, Web Service like interfaces and software libraries have been created. In addition to it, the developer may harness the standard languages (Speech Recognition Grammar Specificication) or the specific ones of each vendor.

The “barge-in” functionality, which is necessary for a natural speech interaction, uses extremely efficient algorithms, that also ensure robustness against background noise.