MnTTS2
收藏资源简介:
MnTTS2是由内蒙古大学开发的开放源代码多说话人蒙古语文本到语音合成数据集。该数据集由三位专业蒙古语播音员录制,每位播音员录制10小时,总计30小时。数据内容涵盖多个主题,如政治、文化、经济、体育等。创建过程中,文本内容经过精心收集和预处理,音频录制在标准录音工作室进行,确保高质量。MnTTS2旨在支持蒙古语TTS研究,特别是用于构建多说话人TTS模型,以满足实际应用需求。
MnTTS2 is an open-source multi-speaker Mongolian text-to-speech (TTS) synthesis dataset developed by Inner Mongolia University. It consists of recordings from three professional Mongolian broadcasters, with each speaker contributing 10 hours of audio, amounting to a total of 30 hours. The dataset covers a wide range of topics including politics, culture, economy, sports and other fields. During its creation, the text materials were carefully collected and preprocessed, and the audio was recorded in a standard recording studio to ensure high data quality. MnTTS2 is intended to support Mongolian TTS research, particularly for the development of multi-speaker TTS models to fulfill practical application requirements.




