Switchboard-1 Release 2

<h3>Introduction</h3><br>
<p>The Switchboard-1 Telephone Speech Corpus (LDC97S62) consists of approximately 260 hours of speech and was originally collected by Texas Instruments in 1990-1, under DARPA sponsorship. The first release of the corpus was published by NIST and distributed by the LDC in 1992-3. Since that release, a number of corrections have been made to the data files as presented on the original CD-ROM set and all copies of the first pressing have been distributed.</p><br>
<p>Switchboard is a collection of about 2,400 two-sided telephone conversations among 543 speakers (302 male, 241 female) from all areas of the United States. A computer-driven robot operator system handled the calls, giving the caller appropriate recorded prompts, selecting and dialing another person (the callee) to take part in a conversation, introducing a topic for discussion and recording the speech from the two subjects into separate channels until the conversation was finished. About 70 topics were provided, of which about 50 were used frequently. Selection of topics and callees was constrained so that: (1) no two speakers would converse together more than once and (2) no one spoke more than once on a given topic.</p><br>
<h3>Data</h3><br>
<p>In this release, assembled and published by the LDC, all known errors affecting the original publication of speech files were corrected. In addition, modifications have been made to the contents of the NIST Sphere headers of all speech files, to identify each file as being part of the new release and to make the usage of the sample_count header field consistent with standard Sphere usage. (In particular, the sample_count field should reflect the number of samples on each channel in the file. In the initial release, this field was improperly set to be the total number of samples in both channels of the file this has been corrected in the new release.)</p><br>
<p>Since the 1997 release, the Switchboard transcripts have been carefully revised at <a href="http://www.isip.piconepress.com/projects/switchboard/" rel="nofollow">The Institute for Signal and Information Processing (ISIP)</a> and additional problems have been discovered and patched. Three speech files, part of the original release, were inadvertently left off the 1997 revision. After corpus users noted some problems in the original speaker attribution table, LDC audited the problem calls and corrected the attributions. The latest version of ISIP transcriptions, the ISIP update of the <a href="http://www.icsi.berkeley.edu" rel="nofollow">ICSI</a> phonetic transcriptions, and corrected word alignments are all available at <a href="http://www.isip.piconepress.com/projects/switchboard/" rel="nofollow">ISIP</a>. The LDC makes the <a href="desc/addenda/swb-multi-annot.summary" rel="nofollow">transcript summaries</a> available via http. Researchers have used SWB-1 data for various annotation projects including discourse annotation/speech acts, part-of-speech tagging and parsing, up-to-date orthographic transcriptions, and phonetic transcriptions. This summary documents which files have been used for the various annotations. In addition to the index of these file characteristics, there is also a table detailing speaker attributes.</p><br>
<h3>Samples</h3><br>
<p>Please view this <a href="desc/addenda/LDC97S62.wav">audio sample</a>.</p><br>
<h3>Updates</h3><br>
<p><strong>08/11/2015:</strong>&nbsp;The three files from the 03/26/2013 update were converted into unshortened&nbsp;sphere. File tables and documentation were updated to reflect the conversion&nbsp;of these files. The corpus is also now available as a web download.&nbsp;All copies of this corpora obtained after the above date include this update.</p><br>
<p><strong>03/26/2013:</strong> Three previously missing files were added to this release. (sw02289.sph, sw04361.sph, sw04379.sph) File tables and documentation were updated to reflect the addition of these files. Please contact <a rel="nofollow">[e-mail]</a> to obtain this update. All copies of this corpora obtained after the above date already include this update.</p><br>
<p><strong>09/29/2011:</strong> Added a file list, available through online docs, to reflect its release on DVD. Also, an updated readme reflects these changes.</p><br>
<p><strong>11/12/2007:</strong> Updated and corrected speaker and call tables are now available online in the corpus documentation directory at&nbsp;<a href="../../../docs/LDC97S62/">https://catalog.ldc.upenn.edu/docs/LDC97S62/</a></p><br>
<p><strong>09/2008:</strong> The Switchboard Dialog Act Corpus is a version of Switchboard-1 Release 2 tagged with a shallow discourse tagset of approximately 60 basic dialog act tags and combinations. The discourse tag-set used is an augmentation of the Discourse Annotation and Markup System of Labeling (DAMSL) tag-set and is referred to as the SWBD-DAMSL labels. These annotations were created in 1997 at the University of Colorado at Boulder, with the goal of building better language models for automatic speech recognition of the Switchboard domain. To that end, the label-set incorporates both traditional sociolinguistic and discourse-theoretic rhetorical relations/adjacency-pairs as well as some more form-based models. This corpus contains labels for 1155 5-minute conversations comprising 205,000 utterances and 1.4 million words. The Switchboard Dialog Act Corpus is available as a&nbsp;<a href="../../../docs/LDC97S62/swb1_dialogact_annot.tar.gz">free download</a>&nbsp;via the&nbsp;<a href="../../../docs/LDC97S62/">online documentation folder</a>.</p><br>
<p>&nbsp;</p><br>
<p>&nbsp;</p></br>
Portions © 1992, 1993, 1997 Trustees of the University of Pennsylvania

<h3>概述</h3><br>
<p>交换台-1电话语音语料库（Switchboard-1 Telephone Speech Corpus, LDC97S62）总时长约260小时，于1990至1991年由德州仪器（Texas Instruments）在美国国防高级研究计划局（DARPA, Defense Advanced Research Projects Agency）资助下首次采集。该语料库首版于1992至1993年由美国国家标准与技术研究院（NIST, National Institute of Standards and Technology）发布、语言数据联盟（LDC, Language Data Consortium）发行。自首版发布以来，针对初始光盘套件中的数据文件已完成多轮修正，且首批印刷的所有副本均已完成分发。</p><br>
<p>交换台语料库包含来自美国全境543名发话者（其中男性302名、女性241名）之间的约2400组双向电话对话。整个流程由计算机驱动的机器人话务系统管控：为发话者播放预设的录音提示音，选择并拨打另一通话对象（受话方）以发起对话，引入讨论主题，并将双方的语音分别录制到独立声道中，直至对话结束。本次语料库共设置约70个讨论主题，其中约50个主题被高频使用。主题与受话方的选取遵循以下约束条件：(1) 任意两名发话者仅可进行一次对话；(2) 任意发话者仅可参与一个主题的对话。</p><br>
<h3>数据</h3><br>
<p>本次由语言数据联盟（LDC）整理发布的版本，已修正所有已知的影响语音文件原始发布的错误。此外，所有语音文件的NIST Sphere文件头内容均已完成修改：一方面用于标识该文件属于本次新版语料库，另一方面使`sample_count`头字段的使用符合标准NIST Sphere规范。（具体而言，`sample_count`字段应反映文件中单个声道的采样点数。在首版发布中，该字段被错误设置为文件双声道的总采样点数，本次新版已修正此问题。）</p><br>
<p>自1997版发布以来，交换台语料库的转写文本已由<a href="http://www.isip.piconepress.com/projects/switchboard/" rel="nofollow">信号与信息处理研究所（ISIP, Institute for Signal and Information Processing）</a>完成细致修订，并修复了额外发现的各类问题。原版发布中的3个语音文件在1997版修订中被意外遗漏。在语料库用户指出原始发话者归属表存在若干问题后，LDC对存在问题的通话进行了审核并修正了归属信息。ISIP的最新版转写文本、<a href="http://www.icsi.berkeley.edu" rel="nofollow">国际计算机科学研究所（ICSI, International Computer Science Institute）</a>语音转写的ISIP修订版，以及修正后的词对齐文件均可在<a href="http://www.isip.piconepress.com/projects/switchboard/" rel="nofollow">ISIP官网</a>获取。LDC还通过HTTP协议提供<a href="desc/addenda/swb-multi-annot.summary" rel="nofollow">转写摘要文档</a>。研究者已将SWB-1（Switchboard-1）语料应用于各类标注任务，包括语篇标注/言语行为标注、词性标注与句法分析、最新正字法转写以及语音转写。本摘要文档将说明各标注任务所使用的具体文件。除上述文件特征索引外，文档还包含一张详述发话者属性的表格。</p><br>
<h3>样本</h3><br>
<p>请查看<a href="desc/addenda/LDC97S62.wav">语音样本</a>。</p><br>
<h3>更新记录</h3><br>
<p><strong>2015年8月11日：</strong>&nbsp;将2013年3月26日更新中涉及的3个文件转换为未压缩的Sphere格式。同步更新了文件列表与文档以反映该转换操作。本语料库现已支持网页下载。2015年8月11日之后获取的所有语料库副本均已包含本次更新。</p><br>
<p><strong>2013年3月26日：</strong>本次版本新增了此前遗漏的3个语音文件（sw02289.sph、sw04361.sph、sw04379.sph）。同步更新了文件列表与文档以反映本次新增文件。可通过联系<a rel="nofollow">[e-mail]</a>获取本次更新。2013年3月26日之后获取的所有语料库副本均已包含本次更新。</p><br>
<p><strong>2011年9月29日：</strong>新增了可通过在线文档获取的文件列表，以适配本语料库的DVD发行版本。同时更新了自述文档以反映上述变更。</p><br>
<p><strong>2007年11月12日：</strong>更新并修正后的发话者与通话列表现已可在线获取，存放于语料库文档目录：<a href="../../../docs/LDC97S62/">https://catalog.ldc.upenn.edu/docs/LDC97S62/</a></p><br>
<p><strong>2008年9月：</strong>交换台对话行为语料库（Switchboard Dialog Act Corpus）是交换台-1第二版的标注版本，采用包含约60种基础对话行为标签及其组合的浅层语篇标记集。本次使用的标记集是对话语标注与标记系统（DAMSL, Discourse Annotation and Markup System of Labeling）标记集的扩展，被称为SWBD-DAMSL标签。该标注工作于1997年由科罗拉多大学博尔德分校完成，旨在为交换台领域的自动语音识别任务构建更优质的语言模型。为此，该标签集融合了传统社会语言学与语篇理论中的修辞关系/毗邻对，以及部分基于形式化的模型。本语料库包含1155段5分钟对话的标注标签，总计20.5万个发话单元、140万个词汇。交换台对话行为语料库可通过<a href="../../../docs/LDC97S62/">在线文档目录</a>以<a href="../../../docs/LDC97S62/swb1_dialogact_annot.tar.gz">免费下载</a>方式获取。</p><br>
<p>&nbsp;</p><br>
<p>&nbsp;</p>
<p>部分内容 © 1992、1993、1997 宾夕法尼亚大学托管委员会</p>

搜集汇总

背景与挑战

背景概述

Switchboard-1 Release 2是一个英语电话对话语音数据集，包含约260小时的录音，涉及543名美国说话者，用于说话人识别和语音识别研究。该数据集是1993年发布的修订版本，通过计算机机器人操作收集，覆盖多种话题，并包含多种语言注释和更新，如对话行为标注和错误纠正。