Tabular Priority Control Design: An Automated Transcription Comparison Framework for Phonetics

Abstract

Phonetics is the study and systematic classification of speech sounds. It is a discipline of linguistics that deals with the science of sound production, transmission, and reception by humans. Phonetic transcription is the visual depiction of speech sounds utilizing written symbols, IPA (International Phonetic Alphabet) being the most common standard. There have been several computational tools in the past for comparing phonetic transcriptions, and some are either difficult to access and integrate into research, or the learning curve is too high for regular users. This thesis describes the development of an online computational tool, called APTct (Automated Phonetic Transcription comparison tool), with unique features, including a comparison visualization module for optimal alignments and scoring operations. The main highlight of this research is creating a unique tabular priority control system designed to control the path taken by the algorithm while performing phonetic transcription comparisons. This system is very versatile in its implementation, and it applies to several other tabular-style and dynamic programming-based algorithms.

APTct underwent rigorous analysis for precise implementation of phonetic and algorithmic principles. Functional testing of results showed a very high degree of accuracy compared with results obtained from an expert’s hand scoring of phonetic transcription comparisons. APTct is an innovative and validated web-based tool, freely accessible to the public. It is available online at the following URL: https://aptct.auburn.edu.