数据堂—10.4小时中文女声客服合成库

10.4小时中文女声客服合成库，由中文母语发音人录制，声音甜美，语料音素覆盖均衡，专业语音学家参与标注，精准匹配语音合成的研发需求

This is a 10.4-hour Chinese female voice customer service speech synthesis corpus. It was recorded by native Chinese female speakers, featuring a sweet and pleasant vocal timbre. The corpus has balanced phoneme coverage, and was professionally annotated by phoneticians, making it precisely tailored to the research and development needs of speech synthesis technologies.

搜集汇总

背景与挑战

背景概述

该数据集是一个10.4小时的中文女声客服合成库，专为语音合成测试设计，包含9,286句客服场景和日常口语的录音，由声音甜美的母语者录制，并提供了字符、音素及韵律标注。数据以48kHz、16bit的WAV单声道格式提供，采用专业录音室环境采集，版权归数据堂所有，属于商业数据。