ChildMandarin: A Comprehensive Mandarin Speech Dataset for Young Children Aged 3-5

TL;DR

This paper introduces ChildMandarin, a comprehensive Mandarin speech dataset for children aged 3-5, enabling improved ASR and speaker verification research for young children's speech in Mandarin.

Contribution

The paper presents a new, large-scale Mandarin speech dataset for young children, with detailed analysis and evaluation of ASR and SV models, addressing a critical resource gap.

Findings

Fine-tuning pre-trained models significantly improves ASR performance.

The dataset supports effective speaker verification despite children's vocal variability.

ASR models trained from scratch show promising results on child speech.

Abstract

Automatic speech recognition (ASR) systems have advanced significantly with models like Whisper, Conformer, and self-supervised frameworks such as Wav2vec 2.0 and HuBERT. However, developing robust ASR models for young children's speech remains challenging due to differences in pronunciation, tone, and pace compared to adult speech. In this paper, we introduce a new Mandarin speech dataset focused on children aged 3 to 5, addressing the scarcity of resources in this area. The dataset comprises 41.25 hours of speech with carefully crafted manual transcriptions, collected from 397 speakers across various provinces in China, with balanced gender representation. We provide a comprehensive analysis of speaker demographics, speech duration distribution and geographic coverage. Additionally, we evaluate ASR performance on models trained from scratch, such as Conformer, as well as fine-tuned pre-trained models like HuBERT and Whisper, where fine-tuning demonstrates significant performance improvements. Furthermore, we assess speaker verification (SV) on our dataset, showing that, despite the challenges posed by the unique vocal characteristics of young children, the dataset effectively supports both ASR and SV tasks. This dataset is a valuable contribution to Mandarin child speech research. The dataset is now open-source and freely available for all academic purposes on https://github.com/flageval-baai/ChildMandarin.