https://community.intel.com/t5/Intel-Integrated-Performance/IIR-filter-diverges/m-p/870156#M8848 Hi,<BR />I am using IPP 5.1. I am trying to implement an IIR, Butterworth with cutoff at 100Hz. I have obtained the coefficients from Matlab and and tried the filter there; I also tried the same coefficients in Scilab and also the filter worked pretty well.<BR />When I try to use those coefficients with IPP and this function ippsIIR_32f, the filter diverges, I beleive I have callled all initialization and allocation methods before calling ippsIIR_32f...... Has anybody have gotten good success with an IIR from IPP ?<BR />Did you have to do any special arrangements onthe coefficient ordering that is not standard?<BR />Thank you.<BR />God bless you<BR /><BR /><BR /> Wed, 11 Apr 2007 17:39:37 GMT srodriguez 2007-04-11T17:39:37Z https://community.intel.com/t5/Intel-Integrated-Performance/IIR-filter-diverges/m-p/870156#M8848 Hi,<BR />I am using IPP 5.1. I am trying to implement an IIR, Butterworth with cutoff at 100Hz. I have obtained the coefficients from Matlab and and tried the filter there; I also tried the same coefficients in Scilab and also the filter worked pretty well.<BR />When I try to use those coefficients with IPP and this function ippsIIR_32f, the filter diverges, I beleive I have callled all initialization and allocation methods before calling ippsIIR_32f...... Has anybody have gotten good success with an IIR from IPP ?<BR />Did you have to do any special arrangements onthe coefficient ordering that is not standard?<BR />Thank you.<BR />God bless you<BR /><BR /><BR /> Wed, 11 Apr 2007 17:39:37 GMT https://community.intel.com/t5/Intel-Integrated-Performance/IIR-filter-diverges/m-p/870156#M8848 srodriguez 2007-04-11T17:39:37Z https://community.intel.com/t5/Intel-Integrated-Performance/IIR-filter-diverges/m-p/870157#M8849 <P>Hello,</P> <P>please see answer from our expert</P><PRE><FONT color="#000080" size="2"><P>/* =======================================</P><P>Order of taps = b0,b1,...,bN,a0,a1,...,aN</P><P>N = order</P><P>======================================== */</P></FONT><FONT face="Arial CYR" color="#000080" size="2"><P>It is mentioned in the manual. Also Butterworth coefficients generation is<BR />available in ipp52 (the only difference with Matlab is that input cutoff freq<BR />should be 1/2F from used in MatLab (it is also described in the manual).</P></FONT></PRE> <P>Regards,<BR /> Vladimir</P> Thu, 12 Apr 2007 12:48:06 GMT https://community.intel.com/t5/Intel-Integrated-Performance/IIR-filter-diverges/m-p/870157#M8849 Vladimir_Dudnik 2007-04-12T12:48:06Z