topic Problem on applying ippsIIRGenLowpass/Highpass in IntelĀ® Integrated Performance Primitives
https://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830636#M5531
<P>Hi Wong, </P><P>Could you provide some clarification about the code: what is the error of the filtered signal?</P><P>For the code, the FFT result is CCS format ( the data are the complex), but ippsIIR use a real filter for the computation. Is this expected?</P><P>Thanks,<BR />Chao</P>Fri, 03 Sep 2010 08:52:57 GMTChao_Y_Intel2010-09-03T08:52:57ZProblem on applying ippsIIRGenLowpass/Highpass
https://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830635#M5530
Hi,<BR /><BR />I am having difficulties on applying ippsIIRGenLowpass/Highpass since there is no sample code online. What I need is to filter a Fourier transformed signal using Butterworth filter. The code below is what I figure out so far by studying Example 6-10 and 6-16 in IPPS manual.<BR /><BR /><BR />//FFT of input signal<BR />Ipp32f buf_FFT_left[len_FFT];<BR />Ipp32f buf_FFT_right[len_FFT];<BR />Ipp64f buf_FFT_left_64f[len_FFT];<BR />Ipp64f buf_FFT_right_64f[len_FFT];<BR />IppsFFTSpec_R_32f* spec;<BR />ippsFFTInitAlloc_R_32f(&spec, 12, IPP_FFT_DIV_INV_BY_N, ippAlgHintNone );<BR />ippsFFTFwd_RToCCS_32f( buf_left_32f, buf_FFT_left, spec, NULL );<BR />ippsFFTFwd_RToCCS_32f( buf_right_32f, buf_FFT_right, spec, NULL );<BR /><BR />//Butterworth filter<BR />ippsConvert_32f64f(buf_FFT_left,buf_FFT_left_64f,len_buf);<BR />ippsConvert_32f64f(buf_FFT_right,buf_FFT_right_64f,len_buf);<BR />int order=4;<BR />Ipp64f rFreq=10*2/RATE;<BR />int tapslen=2*(order+1);<BR />IppsIIRState_64f* pState;<BR />Ipp64f* pTaps=ippsMalloc_64f(tapslen*sizeof(Ipp64f));<BR />ippsIIRGenHighpass_64f(rFreq,0,order,pTaps,ippButterworth);<BR />ippsIIRInitAlloc_64f(&pState,pTaps,order,0);<BR />ippsIIR_64f_I(buf_FFT_left_64f,len_FFT,pState);<BR />ippsIIR_64f_I(buf_FFT_right_64f,len_FFT,pState);<BR />ippsIIRFree_64f(pState);<BR />ippsConvert_64f32f(buf_FFT_left_64f,buf_FFT_left,len_buf);<BR />ippsConvert_64f32f(buf_FFT_right_64f,buf_FFT_right,len_buf);<BR /><BR /><BR />Without the filter, the FFT signal works in my application, but filtered signal do not work. Is anyone have any idea on what is wrong with my code? May be can point me to some sample code or way to display the signal in graph?<BR /><BR />Thank you very much.<BR /><BR />Wong<BR />Thu, 02 Sep 2010 14:05:16 GMThttps://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830635#M5530Wei_Kin_Wong2010-09-02T14:05:16ZProblem on applying ippsIIRGenLowpass/Highpass
https://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830636#M5531
<P>Hi Wong, </P><P>Could you provide some clarification about the code: what is the error of the filtered signal?</P><P>For the code, the FFT result is CCS format ( the data are the complex), but ippsIIR use a real filter for the computation. Is this expected?</P><P>Thanks,<BR />Chao</P>Fri, 03 Sep 2010 08:52:57 GMThttps://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830636#M5531Chao_Y_Intel2010-09-03T08:52:57ZProblem on applying ippsIIRGenLowpass/Highpass
https://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830637#M5532
<BR />To display the data in graph, do you notice IPP demo at \IPP\6.1.6.056\ia32\demo\ippsDemo.exe.<BR />It can demo the IPP single processing function usage. <BR /><BR />Thanks,<BR />ChaoFri, 03 Sep 2010 08:55:05 GMThttps://community.intel.com/t5/Intel-Integrated-Performance/Problem-on-applying-ippsIIRGenLowpass-Highpass/m-p/830637#M5532Chao_Y_Intel2010-09-03T08:55:05Z