Multilinguality, singing synthesis, acoustic emoticons, and other extensions of the Slovak speech synthesizer for SMS reading

After the successful implementation of our speech synthesizer Kempelen in the SMS to voice service of Eurotel, one of the biggest mobile telephone operators in Slovakia, we have found out that several extensions need to be done to make the service even more important for users. In the paper we describe the Slovak speech synthesizer Kempelen, its Hungarian version, we discuss the need of male/female/children voices. We describe a program enabling developers to create a new voice for testing purposes in 10 minutes. We introduce a new feature – speech singing synthesis in SMS to Voice service. Some basic acoustic and signal processing problems of the singing voice synthesis using speech synthesizer are discussed. Last, but not least we introduce a classification of Acousticons (pre-recorded illustrative sounds) which offer much more possibilities, than only acoustical representation of graphical emoticons. The acoustic, aesthetic, psychological, sociological and commercial aspects were taken into account in the classification scheme design.