2005 Spring NIST Rich Transcription (RT-05S) Evaluation Set

<h3>Introduction</h3><br>
<p>2005 Spring NIST Rich Transcription (RT-05S) Conference Meeting Evaluation Set was developed by LDC and NIST (National Institute of Standards and Technology). It contains approximately 78 hours of English meeting speech, reference transcripts and other material used in the <a href="http://www.itl.nist.gov/iad/mig/tests/rt/2005-spring/index.html" rel="nofollow">RT Spring 2005 evaluation</a>. Rich Transcription (RT) is broadly defined as a fusion of speech-to-text (STT) technology and metadata extraction technologies providing the bases for the generation of more usable transcriptions of human-human speech in meetings. LDC has also released <a href="http://catalog.ldc.upenn.edu/LDC2007S11" rel="nofollow">2004 Spring NIST Rich Transcription (RT-04S) Development Data LDC2007S11</a> and <a href="http://catalog.ldc.upenn.edu/LDC2007S12" rel="nofollow">2004 Spring NIST Rich Transcription (RT-04S) Evaluation Data LDC2007S12</a>.</p><br>
<p>RT-05S included the following tasks in the meeting domain:</p><br>
<ul><br>
<li>Speech-To-Text (STT) -convert spoken words into streams of text</li><br>
<li>Speaker Diarization (SPKR) -find the segments of time within a meeting in which each meeting participant is talking</li><br>
<li>Speech Activity Detection (SAD) - detect when someone in a meeting space is talking</li><br>
</ul><br>
<p>Further information about the evaluation is available on the <a href="http://www.itl.nist.gov/iad/mig/tests/rt/2005-spring/index.html" rel="nofollow">RT-05Spring Evaluation Website</a>. Please note the lecture meeting data is not included in this release.</p><br>
<h3>Data Description</h3><br>
<p>The data in this release consists of portions of meeting speech collected between 2001 and 2005 by the IDIAP Research Institutes Augmented Multi-Party Interaction project (AMI), Martigny, Switzerland International Computer Science Institute (ICSI) at University of California, Berkeley Interactive Systems Laboratories (ISL) at Carnegie Mellon University (CMU), Pittsburgh, PA NIST and Virginia Polytechnic Institute and State University (VT), Blacksburg, VA. Each meeting excerpt contains a head-mic recording for each subject and one or more distant microphone recordings.</p><br>
<p>Reference transcripts for the evaluation excerpts were prepared by LDC according to its <a href="http://www.ldc.upenn.edu/Catalog/docs/LDC2011S06/MeetingDataCarefulSpec-V1.2.pdf" rel="nofollow">Meeting Recording Careful Transcription Guidelines</a>. Those specifications are designed to provide an accurate, verbatim (word-for-word) transcription, time-aligned with the audio file and including the identification of additional audio and speech signals with special mark-up.</p><br>
<h3>Samples</h3><br>
<p>For an example of the data contained in this corpus, review this <a href="desc/addenda/LDC2011S06.wav" rel="nofollow">audio sample</a>&nbsp;and <a href="desc/addenda/LDC2011S06.txt" rel="nofollow">transcript sample</a>.</p></br>
Portions © 2011 Trustees of the University of Pennsylvania

<h3>Introduction</h3><br><p>2005年春季NIST富转录（Rich Transcription，RT-05S）会议评测数据集由语言数据联盟（Linguistic Data Consortium，LDC）与美国国家标准与技术研究院（National Institute of Standards and Technology，NIST）联合研发。该数据集包含约78小时的英语会议语音数据、参考转录文本及其他用于<a href="http://www.itl.nist.gov/iad/mig/tests/rt/2005-spring/index.html" rel="nofollow">2005年春季RT评测</a>的相关材料。富转录（Rich Transcription，RT）广义上指语音转文本（Speech-to-Text，STT）技术与元数据提取技术的融合，可为生成更具可用性的人际会议语音转录文本提供基础支撑。LDC还发布了<a href="http://catalog.ldc.upenn.edu/LDC2007S11" rel="nofollow">2004年春季NIST富转录（RT-04S）开发数据集LDC2007S11</a>与<a href="http://catalog.ldc.upenn.edu/LDC2007S12" rel="nofollow">2004年春季NIST富转录（RT-04S）评测数据集LDC2007S12</a>。</p><br><p>RT-05S涵盖会议领域的以下评测任务：</p><br><ul><br><li>语音转文本（Speech-To-Text，STT）：将口语词汇转换为文本流</li><br><li>说话人聚类（Speaker Diarization，SPKR）：识别会议中每位参会者的发言时段</li><br><li>语音活动检测（Speech Activity Detection，SAD）：检测会议空间内的发言行为</li><br></ul><br><p>更多评测相关信息可访问<a href="http://www.itl.nist.gov/iad/mig/tests/rt/2005-spring/index.html" rel="nofollow">RT-05S评测官网</a>。请注意，本次发布版本未包含讲座类会议数据。</p><br><h3>Data Description</h3><br><p>本次发布的数据，源自2001年至2005年间由以下机构采集的部分会议语音：瑞士马蒂尼的IDIAP研究所增强多方交互（Augmented Multi-Party Interaction，AMI）项目、加州大学伯克利分校国际计算机科学研究所（International Computer Science Institute，ICSI）、宾夕法尼亚州匹兹堡的卡内基梅隆大学交互系统实验室（Interactive Systems Laboratories，ISL）、NIST以及弗吉尼亚州布莱克斯堡的弗吉尼亚理工大学（Virginia Polytechnic Institute and State University，VT）。每份会议节选数据均包含每位受试者的头戴式麦克风录音，以及一路或多路远场麦克风录音。</p><br><p>本次评测节选的参考转录文本由LDC依据其<a href="http://www.ldc.upenn.edu/Catalog/docs/LDC2011S06/MeetingDataCarefulSpec-V1.2.pdf" rel="nofollow">会议录音规范转录指南</a>制作完成。该指南旨在生成准确的逐字转录文本，与音频文件实现时间对齐，并通过特殊标记识别额外音频与语音信号。</p><br><h3>Samples</h3><br><p>若需查看本语料库数据示例，可访问<a href="desc/addenda/LDC2011S06.wav" rel="nofollow">音频样例</a>与<a href="desc/addenda/LDC2011S06.txt" rel="nofollow">转录文本样例</a>。</p></br> 部分内容 © 2011 宾夕法尼亚大学校董会

搜集汇总

背景与挑战

背景概述

该数据集是2005年春季NIST Rich Transcription评估集，包含约78小时的英语会议语音、参考转录和评估材料，用于支持语音识别、说话人分割和语音活动检测等任务。数据来源于2001年至2005年多个研究机构收集的会议录音，采用头戴式麦克风和远场麦克风录制，适用于Rich Transcription技术开发。