Samrómur Queries Icelandic Speech 1.0

<h3>Introduction</h3> <p>Samr&oacute;mur Queries Icelandic Speech 1.0 was developed by the <a href="https://lvl.ru.is/">Language and Voice Lab, Reykjavik University&nbsp;</a>in cooperation with&nbsp;<a href="https://almannaromur.is/">Almannar&oacute;mur, Center for Language Technology</a>. The corpus contains 20 hours of Icelandic prompted queries from 3,809 speakers representing 17,475 utterances.</p> <p>This version 1.0 is equivalent to "Samr&oacute;mur Queries Icelandic Speech 21.12" as used by the Language Technology Programme for Icelandic 2019-2023.</p> <h3>Data</h3> <p>Speech data was collected between October 2019 and December 2021 using the&nbsp;<a href="https://samromur.is">Samr&oacute;mur website</a>&nbsp;which displayed prompts to participants. The prompts were mainly from&nbsp;<a href="http://clarin.is/en/resources/gigaword">The Icelandic Gigaword Corpus</a>, which includes text from novels, news, plays, and from a list of location names in Iceland. Additional prompts were taken from&nbsp;<a href="https://www.visindavefur.is/">the Icelandic Web of Science</a>&nbsp;and others were created by combining a name followed by a question. Prompts and speaker metadata are included in the corpus.</p> <p>The audio data is divided into train, dev, and test sets and is presented as flac compressed, single channel, 16 kHz, 16-bit linear PCM.</p> <h3>Samples</h3> <p>Please view this <a href="desc/addenda/LDC2023S05.flac">audio sample (FLAC)</a>.</p> <h3>Updates</h3> <p>None at this time.</p><h3>介绍</h3><p>Samrómur冰岛语查询语音语料库1.0由雷克雅未克大学语言与语音实验室（<a href="https://lvl.ru.is/">Language and Voice Lab, Reykjavik University</a>）与阿尔曼纳罗穆尔语言技术中心（<a href="https://almannaromur.is/">Almannarómur, Center for Language Technology</a>）联合开发。该语料库包含来自3809名发音者的20小时冰岛语提示式查询语音，共计17475条话语。</p><p>此1.0版本等同于2019-2023年冰岛语语言技术计划所使用的“Samrómur冰岛语查询语音语料库21.12”。</p><h3>数据</h3><p>语音数据于2019年10月至2021年12月间通过<a href="https://samromur.is">Samrómur官网</a>收集，该网站会向参与者展示提示文本。提示文本主要来源于<a href="http://clarin.is/en/resources/gigaword">冰岛语千兆词语料库</a>，该语料库涵盖小说、新闻、戏剧文本以及冰岛地名列表。额外的提示文本取自<a href="https://www.visindavefur.is/">冰岛科学网</a>，另有部分提示文本通过将姓名与问句结合的方式生成。语料库中包含提示文本与发音者元数据。</p><p>音频数据被划分为训练集、开发集与测试集，采用FLAC（Free Lossless Audio Codec）压缩格式，单声道，采样率16kHz，位深度16位线性PCM。</p><h3>示例</h3><p>请查看此<a href="desc/addenda/LDC2023S05.flac">音频示例（FLAC）</a>。</p><h3>更新记录</h3><p>暂无更新。</p>

搜集汇总

背景与挑战

背景概述

该数据集是一个冰岛语语音数据集，包含20小时的提示查询语音，来自3,809名说话者的17,475个话语，专门用于说话人识别、说话人验证和语音识别等任务。数据收集于2019年至2021年，通过Samrómur网站收集，提示文本源自冰岛语十亿词库等多种来源，音频以FLAC格式提供，采样率为16 kHz。