数据堂—1,044小时巴西葡萄牙语手机采集语音数据

1,044小时巴西葡萄牙语手机采集语音数据总计1044小时，由2,038名巴西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。1,044小时巴西葡萄牙语手机采集语音数据的文本经过人工校对，准确率高；匹配主流安卓、苹果系统手机

This dataset consists of 1,044 hours of mobile-collected Brazilian Portuguese speech data, recorded by 2,038 native Brazilian speakers. The recording scripts were designed by linguistic experts, covering multiple categories including general scenarios, conversational interactions, in-vehicle scenarios, smart home scenarios and other types, with rich and diverse content. All scripts of this 1,044-hour Brazilian Portuguese mobile-collected speech dataset have been manually proofread to ensure high accuracy. It is compatible with mainstream Android and Apple smartphones.

搜集汇总

背景与挑战

背景概述

该数据集包含1,044小时的巴西葡萄牙语手机采集语音数据，由2,038名母语者录制，覆盖口语、新闻、交互等多种类别，用于测试巴西葡萄牙语语音识别模型。录音格式为16kHz WAV，在安静室内环境中采集，版权归DataTang所有，支持商业使用。