数据堂—178小时中国儿童麦克风语音采集数据

178小时中国儿童麦克风语音采集数据数据是由739名来自中国的儿童通过高保真麦克风采集录制的语音数据，其中男女比例均衡。178小时中国儿童麦克风语音采集数据的录音内容主要来自儿童教科书、儿童故事书、数字，符合儿童语言使用习惯。录音环境为相对安静的室内，文本均经过人工转写，准确率高

This 178-hour Chinese children's speech dataset was collected using high-fidelity microphones, with data recorded by 739 Chinese children and a balanced gender ratio. The recorded content primarily derives from children's textbooks, children's storybooks and digital materials, which aligns with the natural language usage habits of children. All recordings were collected in relatively quiet indoor environments, and all corresponding texts were manually transcribed with high accuracy.

搜集汇总

背景与挑战

背景概述

该数据集包含178小时的中国儿童麦克风语音采集数据，由739名性别平衡的儿童在安静室内环境中录制，内容涵盖教材、故事书和数字，旨在用于中文语音识别模型的测试任务。录音采用高保真设备，数据格式为44.1kHz、16bit的WAV文件，并已进行人工转录，句子准确率达97%。