LAION’s Got Talent

该数据集是一个综合情感语音数据集，包含了覆盖广泛情感、突发声音、主题和内容的配音样本。此外，该数据集经过了筛选，仅包含没有口音的英语句子。规模上，它包含了大约200小时的语音，来自44.4千个样本。该数据集的任务是进行情感文本到语音的合成。

This dataset is a comprehensive emotional speech dataset that encompasses dubbed speech samples covering a wide spectrum of emotions, transient sounds, topics and contents. Furthermore, the dataset has been filtered to exclusively include accent-free English utterances. In terms of scale, it comprises approximately 200 hours of speech material derived from 44.4 thousand individual samples. The core task associated with this dataset is emotional text-to-speech synthesis.