Browse Prior Art Database

Simplifying Internationalisation for IVR Text to Speech Application Development / Deployment

IP.com Disclosure Number: IPCOM000014950D
Original Publication Date: 2001-Nov-16
Included in the Prior Art Database: 2003-Jun-20
Document File: 2 page(s) / 54K

Publishing Venue

IBM

Abstract

Simplifying Internationalisation for IVR Text to Speech Application Development Deployment Disclosed is a system to simplify the development of multi-language speech applications for Interactive Voice Response (IVR) systems with Text to Speech (TTS) and Machine Translation (MT) technologies.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

  Simplifying Internationalisation for IVR Text to Speech Application Development / Deployment

     Disclosed is a system to simplify the development of multi-language speech applications for Interactive Voice Response (IVR) systems with Text to Speech (TTS) and Machine Translation (MT) technologies.

     An IVR consists of one or more computers equipped with telephony hardware, running speech applications where callers hear audio output, in the form of pre-recorded audio or synthesised TTS prompts, and give input, either by pressing keys on the telephone or speech.

     Traditional IVR systems allow simplified internationalisation of applications by virtue of the fact that there are a supplied set of 'system' audio prompts, which have been pre-recorded in a number of languages. Application developers can also create their own categories of segments, and record each segment in multiple languages . Within each category of segments, a segment can share the same segment name as the corresponding segment for another language, allowing the IVR to play the appropriate segment, according to the locale of the machine (or application). For example, the system segment, mon1(month 1), would result in the english audio segment 'January' to be played when the locale is en_GB (British English), or 'Janvier' when the locale is fr_CA (Canadian French).

     TTS is becoming an increasingly important component of IVR / VoIP (Voice over IP) systems, particularly those capable of running VoiceXML applications. TTS-enabled applications that are to be deployed in multiple locales cannot currently be treated the same way as those prompting with voice segments. To achieve the same functionality with TTS, the application writer would need access to some sort of data structure containing the text translations for each TTS message and then set each TTS message, according to the current locale. Whichever method is employed is an extra overhead on developers - and the current trend, with VoiceXML, is to reduce the amount of 'low level' coding required to develop voice-enabled CRM applications. The problem is that when the application is...