FT SPEECH

FT SPEECH是由哥本哈根信息技术大学创建的一个大型丹麦语语音数据集，包含超过1,800小时的议会会议录音，涉及434名发言者。数据集内容丰富，涵盖了九年的议会会议记录，主要用于提升丹麦语在开放访问语音数据中的资源地位。创建过程中，研究人员对原始录音进行了预处理和时间戳校准，确保数据质量。该数据集主要应用于丹麦语自动语音识别研究，旨在解决现有数据集中缺乏自然口语表达的问题。

FT SPEECH is a large-scale Danish speech dataset developed by the IT University of Copenhagen. It contains over 1,800 hours of parliamentary meeting recordings from 434 speakers, with rich content covering nine years of parliamentary proceedings. The dataset is primarily designed to enhance the standing of Danish as a resource within open-access speech data. During its construction, researchers performed preprocessing and timestamp calibration on the original recordings to guarantee data quality. It is mainly utilized for Danish automatic speech recognition research, with the goal of addressing the lack of natural spoken expressions in existing speech datasets.