MulPack

Multiply the elements of two vectors stored in Pack format.

Syntax

Case 1: Not-in-place operation

IppStatus ippsMulPack_32f(const Ipp32f* pSrc1, const Ipp32f* pSrc2, Ipp32f* pDst, int len);

IppStatus ippsMulPack_64f(const Ipp64f* pSrc1, const Ipp64f* pSrc2, Ipp64f* pDst, int len);

IppStatus ippsMulPack_16s_Sfs(const Ipp16s* pSrc1, const Ipp16s* pSrc2, Ipp16s* pDst, int len, int scaleFactor);

Case 2: In-place operation

IppStatus ippsMulPack_32f_I(const Ipp32f* pSrc, Ipp32f* pSrcDst, int len);

IppStatus ippsMulPack_64f_I(const Ipp64f* pSrc, Ipp64f* pSrcDst, int len);

IppStatus ippsMulPack_16s_ISfs(const Ipp16s* pSrc, Ipp16s* pSrcDst, int len, int scaleFactor);

Parameters

pSrc1, pSrc2	Pointers to the vectors whose elements are to be multiplied together.
pDst	Pointer to the destination vector which stores the result of the multiplication pSrc1 [n] * pSrc2[n].
pSrc	Pointer to the vector whose elements are to be multiplied by the elements of pSrcDst in-place.
pSrcDst	Pointer to the source and destination vector (for the in-place operation).
len	Number of elements in the vector.
scaleFactor	Scale factor, refer to Integer Scaling.

Description

The function ippsMulPack is declared in the ipps.h file. This function multiplies the elements of the vector pSrc1 by the elements of the vector pSrc2, and stores the result in pDst.

The in-place flavors ippsMulPack multiply the elements of the vector pSrc by the elements of the vector pSrcDst, and store the result in pSrcDst.

The functions multiply the packed data according to their packed format. The data in Pack packed format include several real values, the rest are complex. Thus, the function performs several real multiplication operations on real elements and complex multiplication operations on complex data. Such kind of packed data multiplication is usually used for signals filtering with the FFT transform when the element-wise multiplication is performed in the frequency domain.

For the functions with the Sfs suffixes scaling is performed in accordance with the scaleFactor value. When the output value exceeds the data range, the result may become saturated.

Example below shows how to use the function ippsMulPack_32f_I.

Return Values

ippStsNoErr	Indicates no error
ippStsNullPtrErr	Indicates an error when the pSrcDst, pDst, pSrc1, pSrc2, or pSrc pointer is NULL.
ippStsSizeErr	Indicates an error when len is less than or equal to 0.

Using the ippsMulPack Function

void mulpack( void ) {

     Ipp32f x[8], X[8], h[8]={1.0f/3,1.0f/3,1.0f/3,0,0,0,0,0}, H[8];

     IppStatus st;

     IppsFFTSpec_R_32f* spec;

     st = ippsFFTInitAlloc_R_32f(&spec, 3, IPP_FFT_DIV_INV_BY_N, 
         ippAlgHintNone);

     ippsSet_32f( 3, x, 8 );

     x[3] = 5;

     st = ippsFFTFwd_RToPack_32f( x, X, spec, NULL );

     st = ippsFFTFwd_RToPack_32f( h, H, spec, NULL );

     ippsMulPack_32f_I( H, X, 8 );

     st = ippsFFTInv_PackToR_32f( X, x, spec, NULL );

     printf_32f("filtered =", x, 8, st );

     ippsFFTFree_R_32f( spec );

Output: 
    filtered =  3.0 3.0 3.0 3.666667 3.666667 3.666667 3.0 3.0

Matlab* analog: 
    >> x=3*ones(1,8); x(4)=5;h=zeros(1,8); h(1:3)=1/3; 
    real(ifft(fft(x).*fft(h)))