数据堂-909小时菲律宾人说英语手机采集语音数据

909小时菲律宾英语手机采集语音数据，菲律宾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；文本经过人工校对，准确率高；909小时菲律宾人说英语手机采集语音数据可应用于语音识别、机器翻译、声纹识别。

This is a 909-hour speech dataset consisting of Filipino-accented English, collected via mobile devices and recorded by local Filipino speakers. The text materials for recording were designed by linguistic experts, covering multiple scenarios including general daily use, interactive scenarios, in-vehicle environments and smart home scenarios. All these texts have been manually proofread to ensure high accuracy. This dataset can be applied to tasks such as speech recognition, machine translation and speaker verification.

搜集汇总

背景与挑战

背景概述

该数据集包含909小时的菲律宾人说英语的语音数据，由2510名母语者通过手机录制，覆盖通用、交互、家居和汽车等多个类别，适用于语音识别和机器翻译等任务。数据格式为16kHz、16bit的单声道WAV文件，具有高准确度。