The CereVoice speech synthesis SDK is a Cross Platform Software Development Kit. CereVoice enables speech developers to add speech output to any piece of software, on any device. The SDK is designed as component to be incorporated into the developer's application. The core CereVoice library can be wrapped in higher-level languages such as Python to ease the development of new applications.

Technical details

The SDK package includes:

• CereVoice Text-to-Speech Engine library, written in ANSI C for portability and speed
• Python wrapper for the CereVoice library
• CereProc's 'Heather' Scottish-accented female voice (additional voices from the CereProc Voice Library are available at extra cost)
• Full documentation of the CereVoice APIs
• Python open source text normalisation system (including a C interface)
• Example applications written in C++ and Python

CereVoice SDK 2.1.1

The CereVoice SDK version 2.1.1 for Windows and Linux is now available. CereVoice 2.1.1 is the fastest, most feature rich CereVoice ever.

New features in CereVoice 2.1.1:

• Faster latency for all applications using the CereVoice text normaliser
• Improved LTS (letter-to-sound) system for pronouncing unknown words
• Support for multi-lingual language databases, lexicons, and LTS
• Additional normaliser functionality covering mathematical operations, numbers, weights and measures
• Improved audio compression performance

Previously added in CereVoice 2.0.0:

• Low memory footprint with embedded voice load functions
• 'Emotional synthesis' capability for Heather and Katherine voices
• Pitch and rate modification library cerevoice_pmod added
• SSML processing module added to text normalisation
• Smoother pitch modification for f0 contour processing
• Lexicon improvements for all English voices
• Dynamic C normaliser library libcerenorm_wrap_shared
• Static Normaliser C library for Windows (Python is not required)
• Normaliser API change - normaliser rules are in a separate library for compatability with alternate languages and locales
• Now compatible with Spanish and Catalan voices