UHV-OTS

UHV-OTS是由Appen开发的一个大规模、高质量的语音数据集，旨在支持商业和学术研究在语音处理领域的数据需求。该数据集通过人机交互的数据标注流程创建，结合机器预标注和全手动审计，确保数据的高质量和多样性。UHV-OTS包含丰富的说话者变异和元信息，如说话者性别、口音以及领域、主题和背景噪音等，适用于多种语言。此外，Appen每年会捐赠一定数量的免费语音数据集，支持学术和开源社区的研究。

UHV-OTS is a large-scale, high-quality speech dataset developed by Appen, designed to address the data requirements of commercial and academic research in the field of speech processing. This dataset is created through a human-machine interactive data annotation workflow, combining machine pre-annotation and full manual auditing to ensure high data quality and diversity. UHV-OTS includes rich speaker variations and metadata such as speaker gender, accent, domain, topic, background noise and more, and is applicable to multiple languages. Additionally, Appen donates a certain number of free speech datasets annually to support research in academic and open-source communities.