The CereVoice Characterful Speech Synthesiser SDK

Abstract

CereProc R Ltd. have recently released a beta version of a commercial unit selection synthesiser featuring XML control of speech style. The system is freely available for academic use and allows fine control of the rendered speech as well as full timings to interface with avatars and other animation.With reference to this system we will discuss current state-of-theart commercial expressive synthesis, and argue that underlying current approaches to sythesis, and current commercial pressures, make it difficult for many systems to create characterful synthesis. We will present how CereProc's approach differs from the industry standard and how we have attempted to maintain and increase the characterfullness of CereVoice's output.We will outline the expressive synthesis markup that is supported by the system, how these are expressed in underlying digital signal processing and selection tags. Finally we will present the concept of second pass synthesis where cues can be manually tweaked to allow direct control of intonation style.

The Romanian speech synthesis (RSS) corpus: Building a high quality HMM-based speech synthesis system using a high sampling rate

“…However, their availability and applicability is limited. For the purpose of this study, a new text processor was developed, based on the Cerevoice development framework (Aylett and Pidcock, 2007). Languagedependent data has been gathered and probabilistic models have been trained; the front-end outputs HTS format labels comprising 53 kinds of contexts (Zen et al, 2007c).…”

From Non-verbal Signals Sequence Mining to Bayesian Networks for Interpersonal Attitudes Expression

“…The character's speech was identical in both videos, and was produced in English with the Cereproc Text-ToSpeech engine [4]. The non-verbal behavior of the recruiter was however dierent in the two videos of every pair.…”

Towards graceful turn management in human-agent interaction for people with cognitive impairments

“…The screen was able to show the game scene, showing the animated virtual agent, "Billie", as well as lists representing the game state. The agent was controlled by the ASAPRealizer software for behavior realization [9]; text-to-speech was provided by a CereVoice [10] component controlled by the realizer, which was able to provide some realizations of paraverbal signals. A directional microphone and a low-cost eyetracker were mounted below the screen.…”