Normalization of Non Standard Words for Kannada Speech Synthesis

The purpose of summary of an article is to facilitate quick and accurate identification of the topic of published document. The objective is to save a prospective reader's time and effort in finding the useful information in a given article. This paper considers the task of text normalization in concatinative Text To Speech (TTS) synthesis for Kannada language. The main focus is to have a single document summarization tool based on statistical approach. This deals on how non standard Kannada words-acronyms, abbreviations, proper names derived from other languages or clutters, phone numbers, decimal numbers, fractions, ordinary numbers, sequence of numbers, money, dates, measures, titles, times and symbols-are preprocessed before passing it to the TTS system as an input. The paper also discusses about the methodology used to normalize the non Kannada text present in the input text to get an equivalent Kannada as output. The method uses a fast lexical analyzer, Jflex to scan the input to find the non standard words in the given input document. INTRODUCTION A TTS synthesizer generates speech from a given text. Although TTS is not yet able to replicate the quality of recorded human speech, it has improved greatly in recent years. There exist different synthesis technologies suitable for different applications. A non-general system could have a limited vocabulary support and limitations in the length of spoken utterances. Multilingual speech processing has become an interesting area to the research community for many years and the field is receiving renewed interest owing to two strong driving forces [1]. Technical advances in speech recognition and synthesis are posing new challenges and opportunities to researchers. For instance, discriminative features are seeing wide application by the speech recognition community, but additional issues arise when using such features in a multilingual setting. Another situation is the apparent convergence of speech recognition and speech synthesis technologies in the form of statistical parametric methodologies. This convergence enables the investigation of new approaches to unified modeling for automatic speech recognition and TTS synthesis as well as cross-lingual speaker adaptation for TTS. The second driving force is the impetus being provided by both government and industry for technologies to break down domestic and international language barriers. Speech signal of an utterance in a language is the only physical event that can be recorded and reproduced. The signal can be further processed in two directions – signal and linguistic processing. During linguistic processing, signals …