数据堂—24小时中文女声通用合成库

中文女声通用合成库，由中文母语发音人录制，声音温柔亲切。涵盖日常口语、有声读物、新闻、广告、客服、电影解说，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。

Universal Chinese Female Speech Synthesis Corpus: Recorded by native Chinese speakers, this corpus exhibits a warm and friendly vocal quality. It encompasses a wide variety of speech scenarios including daily conversational language, audiobooks, news segments, advertisements, customer service dialogues, and movie commentaries. The corpus has a balanced phoneme coverage, and was annotated by professional phoneticians, precisely meeting the R&D requirements for speech synthesis.

搜集汇总

背景与挑战

背景概述

该数据集是一个24小时中文女声通用合成库，主要用于中文女性通用合成模型的测试任务，包含多种语音内容并平衡了语音元素的覆盖，数据格式为48kHz、16bit、wav、单声道。