Text normalization for Croatian speech synthesis

This paper presents text normalization which is an integral part of any text-to-speech (TTS) synthesis system. Text normalization is a set of methods with a task to write non-standard words (NSW) in full expanded form. The algorithms which transform NSW into Croatian text: numbers, dates, times, abbreviations, acronyms and the most common symbols into their expanded form are presented. The whole taxonomy for classification of non-standard words in Croatian language together with rule-based normalization methods combined with a lookup dictionary are proposed. The paper concludes with a discussion on the possible integration of proposed text normalization into the existing text-to-speech synthesis system.