aidatatang_1505zh

1,505小时 中文普通话语音数据集数据时长1505小时，是数据堂中文普通话语音数据库中的一部分。采集区域覆盖全国34个省级行政区域，参与录音人数达6408人，录音内容超30万条口语化句子。经过专业语音校对人员转写标注，并通过严格质量检验，句标注准确率达98%以上，是行业内句准确率的最高标准。（仅支持学术研究，未经允许禁止商用）

This 1,505-hour Mandarin Chinese speech dataset has a total duration of 1,505 hours and is a part of the Datatang Mandarin Chinese Speech Database. Its data collection covers all 34 provincial-level administrative regions across China, involving 6,408 participants and over 300,000 colloquial spoken sentences. All recordings were transcribed and annotated by professional speech proofreaders and passed strict quality inspections, achieving a sentence-level annotation accuracy rate of over 98%, which reaches the highest industry standard for sentence-level annotation accuracy. This dataset is exclusively for academic research purposes, and commercial use is strictly prohibited without prior authorization.

搜集汇总

背景与挑战

背景概述

aidatatang_1505zh是一个1,505小时的中文普通话语音数据集，覆盖全国34个省级行政区域，由6,408人录制超过30万条口语化句子，标注准确率达98%以上，适用于学术研究但禁止商用。