数据堂—134小时马来语手机采集语音数据

134小时马来语手机采集语音数据由156位马来西亚本地人员参与采集，每人约450句。录音环境安静，录音内容包括经济、娱乐、新闻、口语、数字、字母等多个领域。134小时马来语手机采集语音数据所有文本由人工转写，准确率高

This 134-hour Malay speech dataset was collected via mobile devices, with participation from 156 local Malaysian speakers, each contributing approximately 450 utterances. The recordings were made in quiet environments, covering multiple domains including economy, entertainment, news, colloquial speech, numbers and letters. All transcriptions of the dataset were manually completed with high accuracy.

搜集汇总

背景与挑战

背景概述

该数据集包含134小时的马来语手机采集语音数据，专门用于测试马来语语音识别模型。数据由156名马来西亚本地参与者在安静环境中录制，内容覆盖经济、娱乐、新闻等多个领域，所有文本均经过人工转录，确保高准确性。