https://community.intel.com/t5/Intel-Integrated-Performance/ReconstructDCTBlockIntra-Documentation/m-p/942991#M17863 <DIV>Hi,</DIV> <DIV></DIV> <DIV>there is misprint in documentation, you should not use de-quantization after this function because this operation is performed inside of this function.</DIV> <DIV>Documentation will be corrected in the next release</DIV> <DIV></DIV> <DIV>Regards,</DIV> <DIV> Vladimir</DIV> Thu, 26 Aug 2004 13:15:52 GMT Vladimir_Dudnik 2004-08-26T13:15:52Z https://community.intel.com/t5/Intel-Integrated-Performance/ReconstructDCTBlockIntra-Documentation/m-p/942990#M17862 Hello -<BR /><BR />Page 16-9 of the ippiman suggests that Inverse Quantization must be performed after calling ReconstructDCT...(). However, the documentation for ReconstructDCT...() states clearly that it performs inverse quantization.<BR /><BR />I tried performing the additional InvQuant step while attempting to debug my problems with the AANTransposed functions and wow, they really screw up the output.<BR /><BR />So I believe that the graphic on pg. 16-9 has an extraneous step inserted.<BR /><BR />Peter Doege Wed, 25 Aug 2004 02:53:30 GMT https://community.intel.com/t5/Intel-Integrated-Performance/ReconstructDCTBlockIntra-Documentation/m-p/942990#M17862 pdoege 2004-08-25T02:53:30Z https://community.intel.com/t5/Intel-Integrated-Performance/ReconstructDCTBlockIntra-Documentation/m-p/942991#M17863 <DIV>Hi,</DIV> <DIV></DIV> <DIV>there is misprint in documentation, you should not use de-quantization after this function because this operation is performed inside of this function.</DIV> <DIV>Documentation will be corrected in the next release</DIV> <DIV></DIV> <DIV>Regards,</DIV> <DIV> Vladimir</DIV> Thu, 26 Aug 2004 13:15:52 GMT https://community.intel.com/t5/Intel-Integrated-Performance/ReconstructDCTBlockIntra-Documentation/m-p/942991#M17863 Vladimir_Dudnik 2004-08-26T13:15:52Z