CIEMPIESS Light

<h3>Introduction</h3><br>
<p>CIEMPIESS (Corpus de Investigaci&oacute;n en Espa&ntilde;ol de M&eacute;xico del Posgrado de Ingenier&iacute;a El&eacute;ctrica y Servicio Social) Light was developed by the <a href="http://odin.fi-b.unam.mx/profesores/abelherrera/">Speech Processing Laboratory</a> of the Faculty of Engineering at the <a href="http://www.unam.mx/">National Autonomous University of Mexico</a> (UNAM) and consists of approximately 18 hours of Mexican Spanish radio and television speech and associated transcripts. The goal of this work was to create acoustic models for automatic speech recognition. For more information and documentation see the <a href="http://www.CIEMPIESS.org/">CIEMPIESS-UNAM Project website</a>.</p><br>
<p>CIEMPIESS Light is an updated version of CIEMPIESS, released by LDC as <a href="../../../LDC2015S07">LDC2015S07</a>. This "light" version contains speech and transcripts presented in a revised directory structure that allows for use with the&nbsp;<a href="http://kaldi-asr.org/">Kaldi</a> toolkit.</p><br>
<p>CIEMPIESS Balance (<a href="../../../LDC2018S11">LDC2018S11</a>) is a companion to this corpus. When combined, they produce a gender balanced corpus.</p><br>
<p>LDC has released the following data sets in the CIEMPIESS series:</p><br>
<ul><br>
<li>CIEMPIESS (<a href="../../../LDC2015S07">LDC2015S07</a>)</li><br>
<li>CHM150 (<a href="../../../LDC2016S04">LDC2016S04</a>)</li><br>
<li>CIEMPIESS Balance (<a href="../../../LDC2018S11">LDC2018S11</a>)</li><br>
<li>CIEMPIESS Experimentation (<a href="../../../LDC2019S07">LDC2019S07</a>)</li><br>
</ul><br>
<h3>Data</h3><br>
<p>The speech recordings were collected from <a href="http://podcast.unam.mx/">Podcast UNAM</a>, a program created by <a href="https://www.derecho.unam.mx/cultura-juridica/radio.php">Radio-IUS</a>, and <a href="http://mirador.cuaed.unam.mx/">Mirador Universitario</a>, a TV program broadcast by UNAM. They are comprised of spontaneous conversations in Mexican Spanish between a moderator and guests. Approximately 75% of the speakers were male, and 25% of the speakers were female.</p><br>
<p>The audio was recorded in MP3 stereo format, using a 44.1 kHz sample rate and bit-rate of 128 kbps or higher. Only "clean" utterances were selected from the raw data, meaning that the utterances were made by only one person with no background noises, whispers, music, foreign accents, white noise or static. The audio files were converted to 16 kHz, 16-bit PCM flac format for this release.</p><br>
<p>Transcripts are presented as UTF-8 encoded plain text.</p><br>
<h3>Acknowledgements</h3><br>
<p>The authors would like to thank Alejandro V. Mena, Elena Vera, Ang&eacute;lica Guti&eacute;rrez and Beatriz Ancira for their support with the social service program: "Desarrollo de Tecnolog&iacute;as del Habla.&rdquo; They would also like to thank the social service students for their hard work.</p><br>
<h3>Samples</h3><br>
<p>Please view this&nbsp;<a href="desc/addenda/LDC2017S23.flac">audio sample</a> and <a href="desc/addenda/LDC2017S23.txt">transcript sample</a>.</p><br>
<h3>Updates</h3><br>
<p>None at this time.</p></br>
Portions © 2017 Carlos Daniel Hernández Mena, © 2017 Universidad Nacional Autónoma de México, © 2017 Trustees of the University of Pennsylvania