[Paper Review] Use Pronunciation by Analogy for text to speech system in Persian language

Ali Jowharpour, Masha allah abbasi dezfuli+1 more|arXiv (Cornell University)|Jul 24, 2011

Reading and Literacy Development6 references3 citations

TL;DR

This paper proposes a pronunciation-by-analogy (PbA) approach to improve Persian text-to-speech systems by leveraging semantic, grammatical, and phonetic analogies to resolve ambiguities in pronunciation, especially for homographs and silent short vowels. The method achieves improved accuracy in phonetic transcription by reusing known pronunciations of similar words, significantly reducing reliance on exhaustive rule-based systems.

ABSTRACT

The interest in text to speech synthesis increased in the world .text to speech have been developed formany popular languages such as English, Spanish and French and many researches and developmentshave been applied to those languages. Persian on the other hand, has been given little attentioncompared to other languages of similar importance and the research in Persian is still in its infancy.Persian language possess many difficulty and exceptions that increase complexity of text to speechsystems. For example: short vowels is absent in written text or existence of homograph words. in thispaper we propose a new method for persian text to phonetic that base on pronunciations by analogy inwords, semantic relations and grammatical rules for finding proper phonetic. Keywords:PbA, text to speech, Persian language, FPbA

Motivation & Objective

• To address the lack of robust text-to-speech systems for Persian, a language with complex orthographic and phonological exceptions.
• To reduce dependency on rule-based systems by introducing analogy-based pronunciation prediction.
• To improve phonetic transcription accuracy in Persian by exploiting semantic and grammatical relationships between words.
• To handle common challenges such as missing short vowels and homograph ambiguity in written Persian.

Proposed method

• The method uses pronunciation-by-analogy (PbA) to infer the correct pronunciation of a word based on similar words with known pronunciations.
• It identifies candidate analogs using semantic similarity, grammatical role, and phonetic structure.
• The system applies a weighted scoring mechanism to rank potential pronunciations based on analogy strength.
• It leverages existing lexical resources and morphological rules to identify relevant analogs in the lexicon.
• The approach integrates grammatical rules and part-of-speech tagging to refine candidate selections.
• It uses a feature-based comparison of phonetic and morphological patterns to select the most plausible pronunciation.

Experimental results

Research questions

• RQ1 How can pronunciation ambiguity in Persian text-to-speech be reduced using analogical reasoning?
• RQ2 To what extent can semantic and grammatical relations improve phonetic transcription accuracy?
• RQ3 Can analogy-based methods outperform traditional rule-based systems in handling Persian orthographic exceptions?
• RQ4 How effective is the use of phonetic similarity and word class in selecting correct pronunciations for homographs?

Key findings

• The PbA method significantly improves phonetic transcription accuracy by reducing reliance on exhaustive rule sets.
• The system effectively resolves homograph ambiguity by identifying semantically and grammatically similar words with known pronunciations.
• The integration of semantic and grammatical features enhances the reliability of pronunciation predictions.
• The approach demonstrates feasibility in handling silent short vowels and irregular spelling patterns in Persian.
• The method shows promise for low-resource language TTS by minimizing manual rule creation.
• The results suggest that analogy-based systems can be a scalable alternative to rule-based approaches in morphologically complex languages like Persian.