Text to Speech
Pronunciation of Nouns in Text to Speech systems
E.Veera Raghavendra, Lavanya Prahallad IIIT Hyderabad, India raghavendra@iiit.net, lavanyap@cmu.edu
At present most speech synthesis systems use raw text as their input which is understandable from a human point of view but problematic for the machines since the process of converting text to speech is very complex; in this paper we discuss the need for having a specific SSML tag for each “mention” (1st occurrence, 2nd occurrence) of a proper noun in the text or paragraph. We discuss that when a proper noun appears first time in the text, then it is spoken more prominently than its second or third or subsequent occurrence. We highlight the need for incorporating a specific tag in SSML to take care of this mention-case. The SSML format is a compromise between human and machine needs. SSML is often embedded in Voice-XML scripts to drive interactive telephony systems. However, it also may be used alone, such as for creating audio books. The advantage that SSML brings is that the designers of such language generation systems need only understand the basic SSML language and do not need specialist speech synthesis knowledge. Introduction Speech Synthesis Markup Language (SSML) is an XML-based markup language for speech synthesis applications. SSML directs all Text Analysis steps, providing a standard way to control aspects of speech such as pronunciation, acronym expansion, volume, pitch, rate, range, duration, pause, emphasis, etc., across different synthesis-capable platforms. The intended use of SSML is to improve the quality of synthesized content. Different markup elements impact different stages of the synthesis process. The markup may be produced either automatically, for instance via XSLT or CSS3 from an XHTML document, or by human authoring. Markup may be present within a complete SSML document or as part of a fragment embedded in another language, although no interactions with other languages are specified as
E.Veera Raghavendra, Lavanya Prahallad IIIT Hyderabad, India raghavendra@iiit.net, lavanyap@cmu.edu
At present most speech synthesis systems use raw text as their input which is understandable from a human point of view but problematic for the machines since the process of converting text to speech is very complex; in this paper we discuss the need for having a specific SSML tag for each “mention” (1st occurrence, 2nd occurrence) of a proper noun in the text or paragraph. We discuss that when a proper noun appears first time in the text, then it is spoken more prominently than its second or third or subsequent occurrence. We highlight the need for incorporating a specific tag in SSML to take care of this mention-case. The SSML format is a compromise between human and machine needs. SSML is often embedded in Voice-XML scripts to drive interactive telephony systems. However, it also may be used alone, such as for creating audio books. The advantage that SSML brings is that the designers of such language generation systems need only understand the basic SSML language and do not need specialist speech synthesis knowledge. Introduction Speech Synthesis Markup Language (SSML) is an XML-based markup language for speech synthesis applications. SSML directs all Text Analysis steps, providing a standard way to control aspects of speech such as pronunciation, acronym expansion, volume, pitch, rate, range, duration, pause, emphasis, etc., across different synthesis-capable platforms. The intended use of SSML is to improve the quality of synthesized content. Different markup elements impact different stages of the synthesis process. The markup may be produced either automatically, for instance via XSLT or CSS3 from an XHTML document, or by human authoring. Markup may be present within a complete SSML document or as part of a fragment embedded in another language, although no interactions with other languages are specified as