Data Set Description
This data set consists of 50 Cantonese audio recordings that I recorded myself, totaling approximately 3 minutes. The recordings were made using equipment in the school's language lab in a quiet environment. The sentences include various sentence structures such as daily conversations, numbers, and questions.
Problems Encountered and Solutions
Inaccurate Sentence Segmentation: Some sentences had very short pauses, which the automatic markers missed or segmented… See the full description on the dataset page:
https://huggingface.co/datasets/eduhk-compling/11605879_CheungYuLung.
