数据堂—769小时法语手机采集语音数据

769小时法语手机采集语音数据数据总计769小时，由1,623名法国本土发音人参与录制；录音文本由语言专家参与设计，涵盖通用、交互、车载、家居等多类别，内容丰富；文本经过人工校对，准确率高。769小时法语手机采集语音数据同时匹配了主流安卓、苹果系统手机

This dataset includes 769 hours of French speech data collected via mobile phones, recorded by 1,623 native French speakers. The recording scripts were designed by language experts, covering multiple categories such as general scenarios, interactive scenarios, in-vehicle scenarios and smart home scenarios, with rich and diverse content. All scripts have undergone manual proofreading to ensure high accuracy. The data was collected using mainstream Android and Apple smartphones.

搜集汇总

背景与挑战

背景概述

该数据集包含769小时的法语手机采集语音，由1623名法国母语者录制，内容覆盖通用话题、交互场景和命令等多样类别，用于法语语音识别模型的评估任务。数据采用16kHz、16位单声道WAV格式，在安静室内环境下采集。