<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic DCT lifting, Why &amp; How in Intel® Integrated Performance Primitives</title>
    <link>https://community.intel.com/t5/Intel-Integrated-Performance/DCT-lifting-Why-How/m-p/880301#M10117</link>
    <description>I am working on an application to extract MFCC features&lt;BR /&gt;&lt;BR /&gt;Having in mind ippsDCTLifterInitAlloc_Mul_16s(IppsDCTLifterState_16s** pDCTLifter, int lenDCT, const Ipp32f* pLifter, int lenCeps)&lt;BR /&gt;&lt;BR /&gt;there are two somewhat similar variables 'lenDCT' and 'lenCeps', the latter stamds for the amount of features being extracted, I guess that there might be cases where lenDCT is larger then lenCeps (otherwise there would have been only a single arg and not two ), on the aurora samplenumDCT is set to 23, why does more DCT coefficients then desierered features are needed? I guess the answer is mathematical in nature.&lt;BR /&gt;&lt;BR /&gt;Any help would be appreciated.</description>
    <pubDate>Sun, 15 Mar 2009 11:55:45 GMT</pubDate>
    <dc:creator>nadavvolicon_com</dc:creator>
    <dc:date>2009-03-15T11:55:45Z</dc:date>
    <item>
      <title>DCT lifting, Why &amp; How</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/DCT-lifting-Why-How/m-p/880301#M10117</link>
      <description>I am working on an application to extract MFCC features&lt;BR /&gt;&lt;BR /&gt;Having in mind ippsDCTLifterInitAlloc_Mul_16s(IppsDCTLifterState_16s** pDCTLifter, int lenDCT, const Ipp32f* pLifter, int lenCeps)&lt;BR /&gt;&lt;BR /&gt;there are two somewhat similar variables 'lenDCT' and 'lenCeps', the latter stamds for the amount of features being extracted, I guess that there might be cases where lenDCT is larger then lenCeps (otherwise there would have been only a single arg and not two ), on the aurora samplenumDCT is set to 23, why does more DCT coefficients then desierered features are needed? I guess the answer is mathematical in nature.&lt;BR /&gt;&lt;BR /&gt;Any help would be appreciated.</description>
      <pubDate>Sun, 15 Mar 2009 11:55:45 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/DCT-lifting-Why-How/m-p/880301#M10117</guid>
      <dc:creator>nadavvolicon_com</dc:creator>
      <dc:date>2009-03-15T11:55:45Z</dc:date>
    </item>
  </channel>
</rss>

