数据堂—10小时客家方言女声合成库

10小时客家方言女声合成库，由梅州客家发音人录制。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

A 10-hour Hakka dialect female speech synthesis corpus recorded by Hakka speakers from Meizhou. The corpus features well-balanced phoneme coverage, with annotations completed by professional phoneticians, and precisely meets the R&D needs of speech synthesis research and development.

搜集汇总

背景与挑战

背景概述

该数据集是一个10小时的梅州客家方言女声语音合成库，专为语音合成测试任务设计，采用专业录音设备在录音室环境下录制，具备平衡的音素覆盖和精准的标注。数据格式为48kHz、24bit的WAV文件，适用于语音合成研发。