<?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 Re: 1D - FFT in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899396#M11147</link>
    <description>&lt;DIV style="margin:0px;"&gt;&lt;/DIV&gt;
&lt;BR /&gt;Hi,&lt;BR /&gt;&lt;BR /&gt;It depends on what you need from the spectrum y. If you want to discard the phase and use the power, you should use the magnitude (sqrt(re^2+im^2)) of the y's complex numbers.&lt;BR /&gt;Alternatively, if you needa real-to-real transform, look into trigonometric transforms section of MKL Reference Manual.&lt;BR /&gt;&lt;BR /&gt;Dima</description>
    <pubDate>Fri, 20 Feb 2009 04:25:07 GMT</pubDate>
    <dc:creator>Dmitry_B_Intel</dc:creator>
    <dc:date>2009-02-20T04:25:07Z</dc:date>
    <item>
      <title>1D - FFT</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899394#M11145</link>
      <description>Hi, &lt;BR /&gt;&lt;BR /&gt;I am a newbie using the FFT, I have some question about the input and output data format using the FFT Functions. &lt;BR /&gt;&lt;BR /&gt;
&lt;PRE&gt;[cpp]  DFTI_DESCRIPTOR *my_desc_handle;
  long status;
  status = DftiCreateDescriptor( &amp;amp;my_desc_handle, DFTI_SINGLE, DFTI_REAL, 1, N);
  status = DftiCommitDescriptor( my_desc_handle);
  status = DftiComputeForward( my_desc_handle, y);
  status = DftiFreeDescriptor( &amp;amp;my_desc_handle );[/cpp]&lt;/PRE&gt;
&lt;BR /&gt;Well, my input data y is real. After the transformation y contains not only real but also imaginary data. Is there a workaround ?&lt;BR /&gt;&lt;BR /&gt;thx</description>
      <pubDate>Thu, 19 Feb 2009 18:47:49 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899394#M11145</guid>
      <dc:creator>sicb0161</dc:creator>
      <dc:date>2009-02-19T18:47:49Z</dc:date>
    </item>
    <item>
      <title>Re: 1D - FFT</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899395#M11146</link>
      <description>&lt;DIV style="margin:0px;"&gt;&lt;/DIV&gt;
&lt;BR /&gt;Hi,&lt;BR /&gt;&lt;BR /&gt;It depends on what you need from the spectrum y. If you want to discard the phase and use the power, you should use the magnitude (sqrt(re^2+im^2)) of the y's complex numbers.&lt;BR /&gt;Alternatively, if you needa real-to-real transform, look into trigonometric transforms section of MKL Refernce Manual.&lt;BR /&gt;&lt;BR /&gt;Dima</description>
      <pubDate>Fri, 20 Feb 2009 04:24:50 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899395#M11146</guid>
      <dc:creator>Dmitry_B_Intel</dc:creator>
      <dc:date>2009-02-20T04:24:50Z</dc:date>
    </item>
    <item>
      <title>Re: 1D - FFT</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899396#M11147</link>
      <description>&lt;DIV style="margin:0px;"&gt;&lt;/DIV&gt;
&lt;BR /&gt;Hi,&lt;BR /&gt;&lt;BR /&gt;It depends on what you need from the spectrum y. If you want to discard the phase and use the power, you should use the magnitude (sqrt(re^2+im^2)) of the y's complex numbers.&lt;BR /&gt;Alternatively, if you needa real-to-real transform, look into trigonometric transforms section of MKL Reference Manual.&lt;BR /&gt;&lt;BR /&gt;Dima</description>
      <pubDate>Fri, 20 Feb 2009 04:25:07 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899396#M11147</guid>
      <dc:creator>Dmitry_B_Intel</dc:creator>
      <dc:date>2009-02-20T04:25:07Z</dc:date>
    </item>
    <item>
      <title>Re: 1D - FFT</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899397#M11148</link>
      <description>&lt;DIV style="margin:0px;"&gt;
&lt;DIV id="quote_reply" style="width: 100%; margin-top: 5px;"&gt;
&lt;DIV style="margin-left:2px;margin-right:2px;"&gt;Quoting - &lt;A href="https://community.intel.com/en-us/profile/260197"&gt;sicb0161&lt;/A&gt;&lt;/DIV&gt;
&lt;DIV style="background-color:#E5E5E5; padding:5px;border: 1px; border-style: inset;margin-left:2px;margin-right:2px;"&gt;&lt;EM&gt;Hi, &lt;BR /&gt;&lt;BR /&gt;I am a newbie using the FFT, I have some question about the input and output data format using the FFT Functions. &lt;BR /&gt;&lt;BR /&gt;
&lt;/EM&gt;&lt;PRE&gt;&lt;EM&gt;[cpp]  DFTI_DESCRIPTOR *my_desc_handle;&lt;BR /&gt;  long status;&lt;BR /&gt;  status = DftiCreateDescriptor( &amp;amp;my_desc_handle, DFTI_SINGLE, DFTI_REAL, 1, N);&lt;BR /&gt;  status = DftiCommitDescriptor( my_desc_handle);&lt;BR /&gt;  status = DftiComputeForward( my_desc_handle, y);&lt;BR /&gt;  status = DftiFreeDescriptor( &amp;amp;my_desc_handle );[/cpp]&lt;/EM&gt;&lt;/PRE&gt;
&lt;BR /&gt;Well, my input data y is real. After the transformation y contains not only real but also imaginary data. Is there a workaround ?&lt;BR /&gt;&lt;BR /&gt;thx&lt;/DIV&gt;
&lt;/DIV&gt;
&lt;/DIV&gt;
Hi,&lt;BR /&gt;&lt;BR /&gt;if I recall correctly you need to use DFTISetValue and select (from many) options you need, say, something along the lines:&lt;BR /&gt;
&lt;PRE&gt;[cpp]! set packed format for output complex conjugate-symmetric data&lt;BR /&gt;status%flag=DFTISetValue(desc_handle, ....., .....)[/cpp]&lt;/PRE&gt;
where ..... go for named constants. (for example DFTI_REAL_REAL)&lt;BR /&gt;&lt;BR /&gt;Take a look at Named Constants and storage schemes. I believe you'll find many good examples on disk.&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;A.&lt;BR /&gt;</description>
      <pubDate>Fri, 20 Feb 2009 04:27:15 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899397#M11148</guid>
      <dc:creator>ArturGuzik</dc:creator>
      <dc:date>2009-02-20T04:27:15Z</dc:date>
    </item>
    <item>
      <title>Re: 1D - FFT</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899398#M11149</link>
      <description>&lt;DIV style="margin:0px;"&gt;
&lt;DIV id="quote_reply" style="width: 100%; margin-top: 5px;"&gt;
&lt;DIV style="margin-left:2px;margin-right:2px;"&gt;Quoting - &lt;A href="https://community.intel.com/en-us/profile/93647"&gt;Dmitry Baksheev (Intel)&lt;/A&gt;&lt;/DIV&gt;
&lt;DIV style="background-color:#E5E5E5; padding:5px;border: 1px; border-style: inset;margin-left:2px;margin-right:2px;"&gt;&lt;EM&gt; &lt;BR /&gt;Hi,&lt;BR /&gt;&lt;BR /&gt;It depends on what you need from the spectrum y. If you want to discard the phase and use the power, you should use the magnitude (sqrt(re^2+im^2)) of the y's complex numbers.&lt;BR /&gt;Alternatively, if you needa real-to-real transform, look into trigonometric transforms section of MKL Refernce Manual.&lt;BR /&gt;&lt;BR /&gt;Dima&lt;/EM&gt;&lt;/DIV&gt;
&lt;/DIV&gt;
&lt;/DIV&gt;
Thx Dima ,&lt;BR /&gt;&lt;BR /&gt;Hi, yes I could do so, but I only want to have the real part of &lt;STRONG&gt;y&lt;/STRONG&gt;. After computing the FFT, &lt;STRONG&gt;y&lt;/STRONG&gt; contains imaginary and real data. To be more precise :
&lt;PRE&gt;[cpp]y&lt;I&gt; = (real part) if i = even,  
       (imag part) if i = odd.[/cpp]&lt;/I&gt;&lt;/PRE&gt;
But &lt;STRONG&gt;y&lt;/STRONG&gt; is only of length&lt;STRONG&gt; n&lt;/STRONG&gt; , thus only containing half the data. &lt;BR /&gt;&lt;BR /&gt;Thx&lt;BR /&gt;</description>
      <pubDate>Fri, 20 Feb 2009 11:34:37 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899398#M11149</guid>
      <dc:creator>sicb0161</dc:creator>
      <dc:date>2009-02-20T11:34:37Z</dc:date>
    </item>
    <item>
      <title>Re: 1D - FFT</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899399#M11150</link>
      <description>&lt;DIV style="margin:0px;"&gt;&lt;/DIV&gt;
Right, for size=N real-to complex transform the y contains only N/2+1 complex values. The remaining values can be restored using conjugate-even symmetry of y: y[N-n]=conj(y&lt;N&gt;). &lt;BR /&gt;Dima&lt;/N&gt;</description>
      <pubDate>Fri, 20 Feb 2009 15:12:43 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/1D-FFT/m-p/899399#M11150</guid>
      <dc:creator>Dmitry_B_Intel</dc:creator>
      <dc:date>2009-02-20T15:12:43Z</dc:date>
    </item>
  </channel>
</rss>

