Czech Broadcast Conversation Speech

<h3>Introduction</h3> <p> Czech Broadcast Conversation Speech was prepared by researchers at the University of West Bohemia, Pilsen, Czech Republic, and consists of 40 hours of speech recorded from Czech Radio 1 in 2003. Transcripts corresponding to the audio files in this corpus are provided in <a href="http://catalog.ldc.upenn.edu/LDC2009T20" rel="nofollow">Czech Broadcast Conversation MDE Transcripts (LDC2009T20)</a>. These corpora join LDC's other Czech broadcast data sets: <a href="http://catalog.ldc.upenn.edu/LDC2004S01" rel="nofollow">Czech Broadcast News Speech (LDC2004S01)</a>, <a href="http://catalog.ldc.upenn.edu/LDC2004T01" rel="nofollow">Czech Broadcast News Transcripts (LDC2004T01)</a>, <a href="http://catalog.ldc.upenn.edu/LDC2000S89" rel="nofollow">Voice of America (VOA) Czech Broadcast News Audio (LDC2000S89)</a>, and <a href="http://catalog.ldc.upenn.edu/LDC2000T53" rel="nofollow">Voice of America (VOA) Czech Broadcast News Transcripts (LDC2000T53)</a>. </p><p>Czech Broadcast Conversation Speech consists of 72 single channel recordings of Radioforum, a live talk program broadcast by <a href="http://www.rozhlas.cz/radiozurnal/portal/" rel="nofollow">Czech Radio 1 (CRo1)</a> every weekday evening. Its format consists of invited guests (most often politicians) spontaneously answering topical questions posed by one or two interviewers. The number of interviewees in a single program varies from one to three, but typically, one interviewer and two interviewees appear in the program. The material includes passages of interactive dialogue, but longer stretches of monologue-like speech comprise the majority of the collected data. Radioforum also has an interactive segment where listeners call the studio and ask their own questions. That telephony speech was not transcribed in the current release. </p><h3>Data</h3> <p>Individual recordings range from 27 minutes to 36 minutes each. The recordings were collected during the period from February 12, 2003 through June 26, 2003. The signal is mono, sampled at 22.05 kHZ with 16-bit resolution, stored in Windows PCM waveform format. The names of the audio files refer to the broadcast date (rfYYMMDD.wav). </p><p>The table below contains details about the audio files and the transcripts: </p> <table width="25%"> <tr> <td>Number of shows</td> <td> 72 </td> </tr> <tr> <td>Number of word tokens</td> <td> 292.6k </td> </tr> <tr> <td>Number of unique words</td> <td> 30.5k </td> </tr> <tr> <td>Duration of transcribed speech</td> <td> 33.0h </td> </tr> <tr> <td>Total number of speakers</td> <td> 128 </td> </tr> <tr> <td>Male speakers</td> <td> 108 </td> </tr> <tr> <td>Female speakers</td> <td> 20 </td> </tr> </table><h3>Samples</h3> <ul> <li><a href="./desc/addenda/LDC2009S02.wav" rel="nofollow">Speech</a></li> </ul><h3>Sponsorship</h3> <p>The completion of this corpus was facilitated by funding provided by the Ministry of Education of the Czech Republic under projects No. ME909 and 2C06020.</p> </br>
Portions © 2003 Cesky rozhlas 1 Radiozurnal, © 2009 Trustees of the University of Pennsylvania

<h3>简介</h3> <p> 捷克广播对话语音语料库（Czech Broadcast Conversation Speech）由捷克共和国比尔森市西波西米亚大学的研究人员构建，收录了2003年从捷克广播电台1台录制的40小时语音数据。该语料库音频文件对应的转写文本可从<a href="http://catalog.ldc.upenn.edu/LDC2009T20" rel="nofollow">捷克广播对话MDE转写文本（Czech Broadcast Conversation MDE Transcripts，LDC2009T20）</a>获取。该语料库与语言数据联盟（LDC）旗下其他捷克广播类数据集同属系列，包括<a href="http://catalog.ldc.upenn.edu/LDC2004S01" rel="nofollow">捷克广播新闻语音语料库（Czech Broadcast News Speech，LDC2004S01）</a>、<a href="http://catalog.ldc.upenn.edu/LDC2004T01" rel="nofollow">捷克广播新闻转写文本（Czech Broadcast News Transcripts，LDC2004T01）</a>、<a href="http://catalog.ldc.upenn.edu/LDC2000S89" rel="nofollow">美国之音（VOA）捷克广播新闻音频语料库（Voice of America (VOA) Czech Broadcast News Audio，LDC2000S89）</a>以及<a href="http://catalog.ldc.upenn.edu/LDC2000T53" rel="nofollow">美国之音（VOA）捷克广播新闻转写文本（Voice of America (VOA) Czech Broadcast News Transcripts，LDC2000T53）</a>。 </p><p>捷克广播对话语音语料库包含72个单声道录音，均来自捷克广播电台1台（CRo1）每日工作日晚间播出的直播谈话节目《Radioforum》。该节目形式为邀请嘉宾（多为政要）即兴回应1至2位主持人提出的时事议题。单期节目受访嘉宾人数为1至3人，常规配置为1位主持人搭配2位受访嘉宾。语料包含交互式对话片段，但绝大多数收录内容为篇幅较长的类独白语音。此外，《Radioforum》设有听众来电互动环节，听众可致电演播室提出问题，但本版发布未对该电话语音进行转写。 </p><h3>数据说明</h3> <p>单条录音时长介于27分钟至36分钟之间，采集时间为2003年2月12日至2003年6月26日。音频信号为单声道，采样率22.05千赫兹，采样精度16比特，存储格式为Windows PCM波形文件。音频文件名以播出日期命名，格式为rfYYMMDD.wav。 </p><p>下表列出了该语料库音频文件与转写文本的详细信息： </p> <table width="25%"> <tr> <td>节目总数</td> <td> 72 </td> </tr> <tr> <td>词元（Token）总数</td> <td> 292.6k </td> </tr> <tr> <td>唯一词数</td> <td> 30.5k </td> </tr> <tr> <td>转写语音总时长</td> <td> 33.0h </td> </tr> <tr> <td>发言者总人数</td> <td> 128 </td> </tr> <tr> <td>男性发言者</td> <td> 108 </td> </tr> <tr> <td>女性发言者</td> <td> 20 </td> </tr> </table><h3>样本</h3> <ul> <li><a href="./desc/addenda/LDC2009S02.wav" rel="nofollow">语音</a></li> </ul><h3>资助说明</h3> <p>本语料库的完成得益于捷克共和国教育部在ME909与2C06020项目下提供的资金支持。</p> </br>
Portions © 2003 Cesky rozhlas 1 Radiozurnal, © 2009 Trustees of the University of Pennsylvania

搜集汇总

背景与挑战

背景概述

该数据集是捷克广播对话语音库，包含40小时从捷克广播电台1录制的语音数据，录制于2003年，涉及128名说话者，主要用于语音识别和说话人识别研究。其特点包括72个单声道录音、自发对话格式，并与其他捷克广播数据集互补，适用于自然语言处理任务。