CUCHILD: A Large-Scale Cantonese Corpus of Child Speech for Phonology and Articulation Assessment

TL;DR

CUCHILD is a comprehensive Cantonese child speech corpus designed to aid research and clinical assessment of phonology and articulation, including children with speech disorders, through detailed acoustical analysis and potential technological applications.

Contribution

This work introduces the first large-scale Cantonese child speech corpus with detailed design, data collection, and analysis, supporting diverse research and clinical applications.

Findings

Acoustical properties of child speech are characterized.

Corpus includes both typical and disordered speech samples.

Potential for improving speech recognition and disorder detection technologies.

Abstract

This paper describes the design and development of CUCHILD, a large-scale Cantonese corpus of child speech. The corpus contains spoken words collected from 1,986 child speakers aged from 3 to 6 years old. The speech materials include 130 words of 1 to 4 syllables in length. The speakers cover both typically developing (TD) children and children with speech disorder. The intended use of the corpus is to support scientific and clinical research, as well as technology development related to child speech assessment. The design of the corpus, including selection of words, participants recruitment, data acquisition process, and data pre-processing are described in detail. The results of acoustical analysis are presented to illustrate the properties of child speech. Potential applications of the corpus in automatic speech recognition, phonological error detection and speaker diarization are also discussed.