Many embedded applications require high quality prompts. These prompts are traditionally recorded but this approach has major limitations since all possible prompts need to be recorded beforehand. In addition, the recordings cannot handle any dynamic content such as street names for a navigation application. A compromise has been to combine recorded utterances for generic parts of sentences with dynamically
audio generated by a text-to-speech (TTS) engine. This approach has also its limitations as the juxtaposition of recorded prompts and TTS outputs results typically in boundary effects that are noticeable. The combination of RealSpeak Solo and PromptSculptor addresses very effectively these problems by providing a powerful framework for tuning generic phrases and allowing a seamless transition with dynamic content.
PromptSculptor™ is a powerful and very flexible text-to-speech tuning tool for RealSpeak. With its intuitive Graphical User Interface (GUI) it allows a developer to quickly create natural sounding prompts tailored to a specific application context and nearly undistinguishable from recorded prompts. PromptSculptor does not require a deep linguistic knowledge or TTS expertise and significantly reduces application development time and voice talent costs.