数据堂—12.6小时中文男声有声书合成库

12.6小时中文男声有声书合成库，由中文母语发音人录制，霸道总裁人设，声音磁性有魅力，有大家风范，专业语音学家参与标注，精准匹配语音合成的研发需求

A 12.6-hour Mandarin male audiobook synthesis dataset recorded by native Mandarin speakers, featuring a "domineering CEO" character persona. The voice is magnetic and charismatic, exuding an air of refined and authoritative elegance. It was annotated by professional phoneticians, and is precisely tailored to meet the research and development requirements of speech synthesis.

搜集汇总

背景与挑战

背景概述

该数据集是一个12.6小时的中文男声有声书合成库，专为语音合成测试任务设计。它由母语中文男性在专业录音室录制，声音磁性且富有魅力，录音格式为48kHz、24bit的.wav单声道文件，内容涵盖小说或影视对话脚本，并进行了字符-声音和韵律标注。数据集版权归数据堂所有，支持商业使用。