Automatic Speech Recognition Datasets in Cantonese: A Survey and New Dataset

TL;DR

This paper surveys existing Cantonese ASR datasets, introduces a new 73.6-hour dataset called MDCC, and demonstrates its effectiveness through experiments with state-of-the-art models, enhancing Cantonese speech recognition.

Contribution

It provides a comprehensive review of Cantonese ASR datasets, introduces the new MDCC dataset, and shows improved ASR performance using multi-dataset learning.

Findings

MDCC improves ASR accuracy over existing datasets

Multi-dataset learning enhances robustness of Cantonese ASR

State-of-the-art models perform better with the new dataset

Abstract

Automatic speech recognition (ASR) on low resource languages improves the access of linguistic minorities to technological advantages provided by artificial intelligence (AI). In this paper, we address the problem of data scarcity for the Hong Kong Cantonese language by creating a new Cantonese dataset. Our dataset, Multi-Domain Cantonese Corpus (MDCC), consists of 73.6 hours of clean read speech paired with transcripts, collected from Cantonese audiobooks from Hong Kong. It comprises philosophy, politics, education, culture, lifestyle and family domains, covering a wide range of topics. We also review all existing Cantonese datasets and analyze them according to their speech type, data source, total size and availability. We further conduct experiments with Fairseq S2T Transformer, a state-of-the-art ASR model, on the biggest existing dataset, Common Voice zh-HK, and our proposed MDCC, and the results show the effectiveness of our dataset. In addition, we create a powerful and robust Cantonese ASR model by applying multi-dataset learning on MDCC and Common Voice zh-HK.