https://community.intel.com/t5/Intel-Integrated-Performance/cross-arch-jpeg-example/m-p/971435#M20610 It doesn't work well for me. It seems that that encoder passes a raw quantization tables to ippiDCTQuantFwdTableInit_JPEG_8u16u() instead a zig-zag ordered ones. I'm using VC++ ( win32 ). Tue, 10 Feb 2004 04:21:18 GMT Maciej_K_ 2004-02-10T04:21:18Z https://community.intel.com/t5/Intel-Integrated-Performance/cross-arch-jpeg-example/m-p/971435#M20610 It doesn't work well for me. It seems that that encoder passes a raw quantization tables to ippiDCTQuantFwdTableInit_JPEG_8u16u() instead a zig-zag ordered ones. I'm using VC++ ( win32 ). Tue, 10 Feb 2004 04:21:18 GMT https://community.intel.com/t5/Intel-Integrated-Performance/cross-arch-jpeg-example/m-p/971435#M20610 Maciej_K_ 2004-02-10T04:21:18Z https://community.intel.com/t5/Intel-Integrated-Performance/cross-arch-jpeg-example/m-p/971436#M20611 <P>Oh, yes, you are right, it is our mistake. Fortunately, it is easy to fix. You need to change function write_dqt(), file jpgmisc.c, line 673:<BR />instead of<BR /> status = write_byte(quant_tbl[zig_zag_tab_index<I>],bitstream);<BR />you need<BR /> status = write_byte(quant_tbl<I>,bitstream);<BR /></I></I></P> <P>Hope it helps.<BR />Thanks for your report.</P> <P>Regards,<BR /> Vladimir</P> <DIV></DIV> Wed, 11 Feb 2004 01:32:17 GMT https://community.intel.com/t5/Intel-Integrated-Performance/cross-arch-jpeg-example/m-p/971436#M20611 Intel_C_Intel 2004-02-11T01:32:17Z