<?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 MPEG4SP Decoder (example) for PCA issues. in Intel® Integrated Performance Primitives</title>
    <link>https://community.intel.com/t5/Intel-Integrated-Performance/MPEG4SP-Decoder-example-for-PCA-issues/m-p/993773#M22642</link>
    <description>&lt;DIV&gt;
&lt;DIV&gt;Hello,&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;I've taken the Simple Profile MPEG4 decoder sample for PCA (l_ipp-PCA-MPEG4-decode_p_4.1.002.tgz) and attempted to compile it with the Windows version of IPP (for the sake of some initialtests).&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;It compiles with minimal changes (eg. ippiMCReconBlock_Round[On|Off]() is a function under PCA IPP, but a macro under windows IPP. Also there is non portable use of &amp;gt;&amp;gt;, but in this case I believe both compilers use arithmetic shift).&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;However when i test it on a MPEG4-SP bitstream (no QPel, no BVOP, no GMC, ...) there seems to be a positional problem with MCR (motion compensation reconstruction). Its as if the half pixel movements are not correctly executed. When I test the same bitstream on the Windows MPEG4 IPP sample, it plays correctly.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;My questions are:&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;1) Are there any known issues with the PCA MPEG4 decoder sample?&lt;/DIV&gt;
&lt;DIV&gt;2) Are the Windows IPP and PCA IPPAPIs 100% compatible?&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;also&lt;/DIV&gt;
&lt;DIV&gt;3)in the documentation for the enumerated type IPPVC_MC_APX, it says:&lt;/DIV&gt;
&lt;DIV&gt;&lt;FONT face="Courier" size="2"&gt;
&lt;P align="left"&gt;typedef enum _IPPVC_MC_APX{IPPVC_MC_APX_FF = 0x0,IPPVC_MC_APX_FH = 0x4,IPPVC_MC_APX_HF = 0x8,IPPVC_MC_APX_HH = 0x0c}IPPVC_MC_APX;&lt;/P&gt;&lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;
&lt;P align="left"&gt;"The first descriptors &lt;/P&gt;&lt;/FONT&gt;&lt;FONT face="Courier" size="2"&gt;FF, FH, HF, HH &lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;show whether prediction is accurate to full pel (&lt;/FONT&gt;&lt;FONT face="Courier" size="2"&gt;F&lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;) or to half a pel (&lt;/FONT&gt;&lt;FONT face="Courier" size="2"&gt;H&lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;). The first letter indicates the accuracy in horizontal direction and the second - in vertical direction."&lt;/FONT&gt;&lt;P&gt;&lt;/P&gt;
&lt;P align="left"&gt;&lt;FONT size="3"&gt;Yet in ippalign.h, the following macro is defined:&lt;/FONT&gt;&lt;/P&gt;
&lt;P align="left"&gt;#define ippiMCReconBlock_RoundOff(pSrc, srcStep, pSrcResidue, pDst, dstStep,&lt;BR /&gt; predictType)&lt;BR /&gt; ippiMC8x8_8u_C1(pSrc, srcStep, pSrcResidue, 16, pDst, dstStep,&lt;BR /&gt; predictType==IPP_VIDEO_INTEGER_PIXEL ? IPPVC_MC_APX_FF:&lt;BR /&gt; (predictType==IPP_VIDEO_HALF_PIXEL_X ? IPPVC_MC_APX_FH:&lt;BR /&gt; (predictType==IPP_VIDEO_HALF_PIXEL_Y ? IPPVC_MC_APX_HF:&lt;BR /&gt; IPPVC_MC_APX_HH)), ippRndZero)&lt;/P&gt;
&lt;P align="left"&gt;Given the documentation shouldnt FH and HF be reversed?&lt;/P&gt;
&lt;P align="left"&gt;(When I try reversing them, I still get motion compensation corruption, so It doesnt solve my problem).&lt;/P&gt;
&lt;P align="left"&gt;Regards,&lt;/P&gt;
&lt;P align="left"&gt;Sam.&lt;/P&gt;
&lt;P align="left"&gt;&lt;/P&gt;&lt;/DIV&gt;&lt;/DIV&gt;</description>
    <pubDate>Mon, 21 Mar 2005 13:13:20 GMT</pubDate>
    <dc:creator>Deleted_U_Intel</dc:creator>
    <dc:date>2005-03-21T13:13:20Z</dc:date>
    <item>
      <title>MPEG4SP Decoder (example) for PCA issues.</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/MPEG4SP-Decoder-example-for-PCA-issues/m-p/993773#M22642</link>
      <description>&lt;DIV&gt;
&lt;DIV&gt;Hello,&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;I've taken the Simple Profile MPEG4 decoder sample for PCA (l_ipp-PCA-MPEG4-decode_p_4.1.002.tgz) and attempted to compile it with the Windows version of IPP (for the sake of some initialtests).&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;It compiles with minimal changes (eg. ippiMCReconBlock_Round[On|Off]() is a function under PCA IPP, but a macro under windows IPP. Also there is non portable use of &amp;gt;&amp;gt;, but in this case I believe both compilers use arithmetic shift).&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;However when i test it on a MPEG4-SP bitstream (no QPel, no BVOP, no GMC, ...) there seems to be a positional problem with MCR (motion compensation reconstruction). Its as if the half pixel movements are not correctly executed. When I test the same bitstream on the Windows MPEG4 IPP sample, it plays correctly.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;My questions are:&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;1) Are there any known issues with the PCA MPEG4 decoder sample?&lt;/DIV&gt;
&lt;DIV&gt;2) Are the Windows IPP and PCA IPPAPIs 100% compatible?&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;also&lt;/DIV&gt;
&lt;DIV&gt;3)in the documentation for the enumerated type IPPVC_MC_APX, it says:&lt;/DIV&gt;
&lt;DIV&gt;&lt;FONT face="Courier" size="2"&gt;
&lt;P align="left"&gt;typedef enum _IPPVC_MC_APX{IPPVC_MC_APX_FF = 0x0,IPPVC_MC_APX_FH = 0x4,IPPVC_MC_APX_HF = 0x8,IPPVC_MC_APX_HH = 0x0c}IPPVC_MC_APX;&lt;/P&gt;&lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;
&lt;P align="left"&gt;"The first descriptors &lt;/P&gt;&lt;/FONT&gt;&lt;FONT face="Courier" size="2"&gt;FF, FH, HF, HH &lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;show whether prediction is accurate to full pel (&lt;/FONT&gt;&lt;FONT face="Courier" size="2"&gt;F&lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;) or to half a pel (&lt;/FONT&gt;&lt;FONT face="Courier" size="2"&gt;H&lt;/FONT&gt;&lt;FONT face="TimesNewRomanPSMT" size="3"&gt;). The first letter indicates the accuracy in horizontal direction and the second - in vertical direction."&lt;/FONT&gt;&lt;P&gt;&lt;/P&gt;
&lt;P align="left"&gt;&lt;FONT size="3"&gt;Yet in ippalign.h, the following macro is defined:&lt;/FONT&gt;&lt;/P&gt;
&lt;P align="left"&gt;#define ippiMCReconBlock_RoundOff(pSrc, srcStep, pSrcResidue, pDst, dstStep,&lt;BR /&gt; predictType)&lt;BR /&gt; ippiMC8x8_8u_C1(pSrc, srcStep, pSrcResidue, 16, pDst, dstStep,&lt;BR /&gt; predictType==IPP_VIDEO_INTEGER_PIXEL ? IPPVC_MC_APX_FF:&lt;BR /&gt; (predictType==IPP_VIDEO_HALF_PIXEL_X ? IPPVC_MC_APX_FH:&lt;BR /&gt; (predictType==IPP_VIDEO_HALF_PIXEL_Y ? IPPVC_MC_APX_HF:&lt;BR /&gt; IPPVC_MC_APX_HH)), ippRndZero)&lt;/P&gt;
&lt;P align="left"&gt;Given the documentation shouldnt FH and HF be reversed?&lt;/P&gt;
&lt;P align="left"&gt;(When I try reversing them, I still get motion compensation corruption, so It doesnt solve my problem).&lt;/P&gt;
&lt;P align="left"&gt;Regards,&lt;/P&gt;
&lt;P align="left"&gt;Sam.&lt;/P&gt;
&lt;P align="left"&gt;&lt;/P&gt;&lt;/DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Mon, 21 Mar 2005 13:13:20 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/MPEG4SP-Decoder-example-for-PCA-issues/m-p/993773#M22642</guid>
      <dc:creator>Deleted_U_Intel</dc:creator>
      <dc:date>2005-03-21T13:13:20Z</dc:date>
    </item>
    <item>
      <title>Re: MPEG4SP Decoder (example) for PCA issues.</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/MPEG4SP-Decoder-example-for-PCA-issues/m-p/993774#M22643</link>
      <description>&lt;DIV&gt;Hi,&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;if you detected any issues with this sample you can submit it to technical support.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;Regards,&lt;/DIV&gt;
&lt;DIV&gt; Vladimir&lt;/DIV&gt;</description>
      <pubDate>Wed, 23 Mar 2005 05:17:57 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/MPEG4SP-Decoder-example-for-PCA-issues/m-p/993774#M22643</guid>
      <dc:creator>Vladimir_Dudnik</dc:creator>
      <dc:date>2005-03-23T05:17:57Z</dc:date>
    </item>
  </channel>
</rss>

