<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<p class="MsoNormal" style="text-align: center;" align="center">LDC2007S12<br>
<a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2007S12"><b>2004
Spring NIST Rich Transcription (RT-04S) Evaluation Data</b></a><br>
<br>
LDC2007T19<br>
<a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2007T19"><b>MITRE
1997 Mandarin Broadcast News Speech Translations(Hub-4NE)</b></a><br>
<br>
</p>
<p class="MsoNormal" style="text-align: center;" align="center"><b>The
:Linguistic Data Consortium (LDC) is pleased to announce the
availability of two new publications.</b><br>
<b><br>
</b></p>
<hr size="2" width="100%">
<p class="MsoNormal" style="margin-bottom: 12pt; text-align: center;"
align="center"><b>New
Publications<br>
</b></p>
<p>(1) <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2007S12">2004
Spring NIST Rich Transcription (RT-04S) Evaluation Data</a> contains
the test
material (meeting speech and reference transcripts) used in the RT-04S
evaluation administered by the <a href="http://www.nist.gov/speech">NIST
(National Institute of Standards and Technology) Speech Group</a>. Rich
Transcription (RT) is broadly defined as a fusion of speech-to-text
technology
and metadata extraction technologies designed to provide the basis for
a
generation of more usable transcriptions of human-human meeting speech.<o:p></o:p></p>
<p>The data in this release consists of portions of meeting speech
collected
and/or transcribed by the International Computer Science Institute
(ICSI) at <st1:city><st1:place>Berkeley</st1:place></st1:city>,
the Interactive Systems Laboratories (ISL) at <st1:place><st1:placename>Carnegie</st1:placename>
<st1:placename>Mellon</st1:placename> <st1:placetype>University</st1:placetype></st1:place>,
NIST and LDC. The complete meeting speech and corresponding transcript
data
sets are available from LDC's catalog as follows: <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004S02">ICSI
Meeting Speech (LDC2004S02)</a>, <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004T04">ICSI
Meeting Transcripts (LDC2004T04)</a>, <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004S05">ISL
Meeting Speech Part 1 (LDC2004S05)</a>, <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004T10">ISL
Meeting Transcripts Part 1 (LDC2004T10)</a>, <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004S09">NIST
Meeting Pilot Corpus Speech (LDC2004S09)</a> and <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004T13">NIST
Meeting Pilot Corpus Transcripts and Metadata (LDC2004T13)</a>.<o:p></o:p></p>
<p>RT-04S included the following tasks in the meeting domain: <o:p></o:p></p>
<dl>
<dt><strong>Speech-to-Text Transcription (STT) tasks</strong> </dt>
<dd><strong>Microphone conditions:</strong>
<ul>
<li>Multiple distant microphones </li>
<li>Single distant microphone </li>
<li>Individual head microphone </li>
</ul>
</dd>
<dd><strong>Processing time conditions:</strong>
<ul>
<li>Unlimited time STT </li>
<li>Less than or equal to twenty times realtime </li>
<li>Less than or equal to ten times realtime </li>
<li>Less than or equal to one times realtime </li>
</ul>
</dd>
<dt><strong>Diarization (SPKR) task (?who spoke when?)</strong> </dt>
<dd><strong>Microphone conditions:</strong>
<ul>
<li>Multiple distant microphones </li>
<li>Single distant microphone </li>
</ul>
</dd>
<dd><strong>Input conditions:</strong>
<ul>
<li>Speech input only </li>
<li>Speech plus reference transcript input </li>
</ul>
</dd>
<dd><strong>Processing time conditions:</strong>
<ul>
<li>Unlimited time </li>
<li>Less than or equal to twenty times realtime </li>
<li>Less than or equal to ten times realtime </li>
<li>Less than or equal to one time realtime</li>
</ul>
</dd>
</dl>
<p class="MsoNormal" style="margin-left: 1in; text-indent: -0.25in;"><o:p></o:p></p>
<p class="MsoNormal" style="margin-bottom: 12pt;">Further information
about the evaluation is
available on the
<a href="http://www.nist.gov/speech/tests/rt/rt2004/spring/">RT-04
Spring
Evaluation Website</a>. <o:p></o:p></p>
<p class="MsoNormal" style="text-align: center;" align="center">*<o:p></o:p></p>
<p class="MsoNormal"><b><br>
</b>(2) <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2007T19">MITRE
1997 Mandarin Broadcast News Transcripts Translations (Hub-4NE)</a> was
developed by The MITRE Corporation and contains segment-aligned English
translations of the 1997 DARPA HUB4-NE Mandarin transcripts. The
original
transcripts and the corresponding broadcast news audio are available as
separate LDC publications, <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC98T24"><span
style="text-decoration: none;">1997 Mandarin Broadcast News
Transcripts (HUB4-NE) (LDC98T24)</span></a> and <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC98S73"><span
style="text-decoration: none;">1997 Mandarin Broadcast News
Speech (HUB4-NE) (LDC98S73)</span></a>. <o:p></o:p></p>
<p>The source data is comprised of 30 hours of recorded Mandarin
broadcasts
collected by the LDC in 1997 from Voice of America, China Central TV
and
KAZN-AM, a commercial radio station based in <st1:place><st1:city>Los
Angeles</st1:city>, <st1:state>CA</st1:state></st1:place>. The
original transcript segmentation is
suitable for speech recognition, but does not support machine
translation and
machine translation evaluation. Therefore, the Mandarin side of these
aligned
transcripts was resegmented for this release; in all other respects,
the
Mandarin transcripts in this publication are identical to the original
transcripts. <o:p></o:p></p>
<p>The dataset in this release consists of 376K words of English text
and 517K
characters of Mandarin text. The English text was produced by
translators with
no access to the original audio. The translators were given specific
guidelines
for translation, and those are included in this distribution. A portion
of the
source data (6%) was translated four times in order to support
experiments in
translation evaluation. <o:p></o:p></p>
<hr size="2" width="100%"><br>
<div align="center"><small><font face="Courier New, Courier, monospace"><br>
Ilya
Ahtaridis<br>
Membership Coordinator</font></small><br>
--------------------------------------------------------------------
<font face="Courier New, Courier, monospace"><br>
</font></div>
<div align="center">
<pre class="moz-signature" cols="72"><b><small><font
face="Courier New, Courier, monospace">
</font></small>Linguistic Data Consortium Phone: (215) 573-1275
University of Pennsylvania Fax: (215) 573-2175
3600 Market St., Suite 810 <a
class="moz-txt-link-abbreviated" href="mailto:ldc@ldc.upenn.edu">ldc@ldc.upenn.edu</a>
Philadelphia, PA 19104 USA <a
class="moz-txt-link-freetext" href="http://www.ldc.upenn.edu">http://www.ldc.upenn.edu</a></b></pre>
</div>
</body>
</html>