ScanSoft Introduces Steffi, a New Female German Text-to-Speech Voice

Latest RealSpeak Voice Provides Superior Voice Quality with Multilingual Capabilities and Extends Company's Unmatched Text-to-Speech Portfolio

London, UK, Voice World Europe, May 6, 2004 - ScanSoft, Inc. (Nasdaq: SSFT), the leading supplier of speech and imaging solutions, today announced the availability of "Steffi," a new female text-to-speech (TTS) voice for RealSpeak™ and RealSpeak™ Solo. Steffi is the first in a series of new TTS voices to be introduced by the SpeechWorks Division of ScanSoft to support telephony-based and embedded applications for European and Asia-Pacific markets.

The SpeechWorks Division offers the industry's preeminent TTS solutions for converting text to human-sounding synthesized speech. Its high-performance TTS solutions offer an unprecedented 23 languages and dialects with a variety of female and male voices to provide customers worldwide with the most comprehensive and flexible solutions available today. RealSpeak TTS has proven itself as a highly scalable solution within a wide range of telephony-based applications for leading organizations including the FIFA World Cup, Nortel Networks and Ericsson. RealSpeak Solo powers TTS solutions for customers including Ford Motor Company and NCR, and educational programs from Auralog, Cornelsen, Oxford University Press and TextHelp.

"Our TTS solutions continue to grow in popularity as the most advanced and natural-sounding TTS solutions available for call center applications and embedded applications such as voice navigation," said Steve Chambers, president of SpeechWorks, a division of ScanSoft. "With large speech technology R&D centers located in Ghent and Aachen, an unmatched portfolio of field-proven speech technologies and a global professional services organization, we are well-positioned to provide superior TTS solutions to customers around the world."

RealSpeak and RealSpeak Solo Features and Applications

RealSpeak™ Steffi is available today for telephony-based applications and through its RealSpeak Solo offering for embedded applications. Regardless of the memory footprint, Steffi delivers premium voice quality and natural-sounding synthesized speech. Steffi also offers advanced multilingual capabilities to properly pronounce foreign, borrowed words from certain languages that are frequently used in the colloquial German language. For example, conversational German often incorporates popular French and English words such as "ordinateur" and "world wide web." In addition, multilingual capabilities can be used to read English song titles and street names in an embedded application. Steffi can be incorporated into a variety of applications including the following:

• Large-scale deployments of sophisticated telephony applications that deliver dynamic, real-time or personal information, including personal account information, email by phone for unified messaging systems, locator services for stores and ATMs, traffic and road conditions, weather reports, automated voice dialing, SMS-to-voice, directory assistance, all kinds of announcement systems and many others.
  
  
• SpeechWorks® TTS solutions can be used to decrease costs by reducing the need for professionally recorded prompts in telephony-based speech applications, resulting in additional savings in application cost and deployment time.
  
  
• Optimized for embedded navigation applications, RealSpeak Solo provides accurate pronunciation of navigation related text commonly used in automotive solutions and navigation devices.
  
  
• RealSpeak Solo supports a range of low-footprint embedded applications designed for optimal performance in talking dictionaries, mobile devices, language learning solutions, gaming consoles, automated teller machines (ATMs) and desktop solutions.

TTS Product Portfolio

The SpeechWorks Division offers a full suite of market-leading TTS solutions for embedded and network-based TTS applications that support open, industry standards such as SSML, MRCP and SAPI for empowering VoiceXML- and SALT-based applications among others. The company's network-based TTS solutions provide customers with the most comprehensive and flexible solutions available today. Its embedded TTS solutions are ideal in mobile, automotive, game and PC/multi-media environments that require memory footprints ranging from 100KB to 80MB. Customers can also take advantage of SpeechWorks' Virtuoso™ Custom Voice Program to create a custom voice for their embedded or network-based TTS application, which can also be specifically tuned for a dedicated application domain, such as navigation. For more information and to sample SpeechWorks' full portfolio of TTS voices, visit www.scansoft.com/products/.

About The SpeechWorks Division

SpeechWorks is the most trusted name in speech. Through its global partner network, SpeechWorks delivers field-proven applications that revolutionize the way people interact. Today, thousands of companies and millions of users around the world depend on SpeechWorks® solutions from ScanSoft to deliver vital business information and simplify every day life. For information, please visit www.scansoft.com/speech.

About ScanSoft, Inc.

ScanSoft, Inc. (Nasdaq: SSFT) is a leading supplier of speech and imaging solutions that are used to automate a wide range of manual processes - saving time, increasing worker productivity and improving customer service. For more information regarding ScanSoft products and technologies, please visit www.ScanSoft.com.

ScanSoft, RealSpeak, and the ScanSoft logo are trademarks or registered trademarks of ScanSoft, Inc. in the United States and other countries. SpeechWorks and Virtuoso are trademarks or registered trademarks of SpeechWorks International, Inc. in the United States and other countries. All other company names or product names may be the trademarks of their respective owners.

Contact Information