Evaluates sum of absolute transformed differences between current and reference 16X8 blocks using 4x4 transform.
Syntax
IppStatus ippiSATD16x8_8u32s_C1R(const Ipp8u* pSrcCur, int srcCurStep, const Ipp8u* pSrcRef, int srcRefStep, Ipp32s* pSATD);
Parameters
pSrcCur |
Pointer to a 16x8 block in the source plane. |
srcCurStep |
Distance in bytes between starts of the consecutive lines in the source image. |
pSrcRef |
Pointer to a 16x8 block in the reference plane. |
scrRefStep |
Distance in bytes between starts of the consecutive lines in the reference image. |
pSATD |
Pointer to the SATD value. |
Description
This function is declared in the ippvc.h header file. The functions ippiSATD16x8_8u32s and ippiSATD16x8_16u32s_C1R evaluate the sum of absolute tranformed differences of all the elements in the current 16x8 block and the corresponding elements in the reference 16x8 block using a 4x4 transform. The transform matrix is as follows:
This matrix is applied to each 4x4 subblock of the 16x8 block.
If D is a 16x8 block of difference, the result of the function is the sum of absolute values of elements of T *D*T for each subblock. The result is stored in pSATD.
This function is used in the H.264 encoder included into Intel IPP Samples. See introduction to Motion Estimation.
Return Values
ippStsNoErr |
Indicates no error. |
ippStsNullPtrErr |
Indicates an error when at least one input pointer is NULL. |