This dataset was recorded in both quiet and noisy environments, with 5226 speakers participating, including 2352 males and 2874 females. All speakers involved in the recording were professionally selected to ensure standard pronunciation and clear enunciation. The recorded text covers information such as news, daily conversations, and SMS messages.
URL… See the full description on the dataset page: https://huggingface.co/datasets/DataoceanAI/Dolphin_Model_Chinese-Taiwanese-Speech-Recognition-Corpus.
