Reconstruct 4X4 inter chroma macroblock for high profile.
IppStatus ippiReconstructChromaInter4x4MB_H264_16s8u_P2R(Ipp16s** ppSrcDstCoeff, Ipp8u* pSrcDstUPlane, Ipp8u* pSrcDstVPlane, Ipp32u srcDstUVStep, Ipp32u cbp4x4, Ipp32s ChromaQPU, Ipp32u ChromaQPV, const Ipp16s* pQuantTableU, const Ipp16s* pQuantTableV, Ipp8u bypassFlag);
IppStatus ippiReconstructChromaInter4x4MB_H264_16s8u_C2R(Ipp16s** ppSrcDstCoeff, Ipp8u* pSrcDstUVPlane, Ipp32u srcDstUVStep, Ipp32u cbp4x4, Ipp32s ChromaQPU, Ipp32u ChromaQPV, const Ipp16s* pQuantTableU, const Ipp16s* pQuantTableV, Ipp8u bypassFlag);
IppStatus ippiReconstructChromaInter4x4_H264High_32s16u_IP2R(const IppiReconstructHighMB_32s16u* pReconstructInfo[2]);
ppSrcDstCoeff |
Pointer to the order of 4x4 blocks of residual coefficients for this macroblock, which are taken as a result of Huffman decoding (2x2 DC U-block, 2x2 DC V-block, 4x4 AC U-blocks, 4x4 AC V-blocks if the block is not zero-filled) in the same order as is shown in Figure 8‑7 of [ITUH264]. Pointer is updated by the function and points to the blocks for the next macroblock. |
pSrcDstUPlane |
Pointer to a macroblock that is reconstructed in the current U plane. This macroblock must contain inter prediction samples. |
pSrcDstVPlane |
Pointer to a macroblock that is reconstructed in the current V plane. This macroblock must contain inter prediction samples. |
pSrcDstUVPlane |
Pointer to a macroblock that is reconstructed in the current UV plane. |
srcDstUVStep |
Plane step. |
cbp4x4 |
Coded block pattern. If cbp4x4 & (1<<(IPPVC_CBP_1ST_CHROMA_DC_BITPOS) is not equal to 0, 2x2 DC U-block is not zero-filled and exists in ppSrcDstCoeff. If cbp4x4 & (1<<(IPPVC_CBP_1ST_CHROMA_DC_BITPOS+ 1)) is not equal to 0, 2x2 DC U-block is not zero-filled and exists in ppSrcDstCoeff. If cbp4x4 & (1<<(IPPVC_CBP_1ST_CHROMA_AC_BITPOS+ i)) is not equal to 0, (0 ≤ i < 4), i-th 4x4 AC U-block is not zero-filled and exists in ppSrcDstCoeff. If cbp4x4 & (1<<(IPPVC_CBP_1ST_CHROMA_AC_BITPOS+ i+4)) is not equal to 0, (0 ≤ i < 4), i-th 4x4 AC U-block is not zero-filled and exists in ppSrcDstCoeff. |
ChromaQPU |
Chroma quantizer for U plane (QPC in [ITUH264]). It must be within the range [0;39]. |
ChromaQPV |
Chroma quantizer for V plane (QPC in [ITUH264]). It must be within the range [0;39]. |
pQuantTableU |
Pointer to the quantization table for U plane (LevelScale(qP%6, i, j ) in [ITUH264]). |
pQuantTableV |
Pointer to the quantization table for V plane (LevelScale(qP%6, i, j ) in [ITUH264]). |
bypassFlag |
Flag enabling lossless coding. |
pReconstructInfo[2] |
Pointer to the IppiReconstructHighMB_32s16u structure. The first pointer pReconstructInfo[0] points to the structure with parameters for restructuring chroma macroblock of plane U, and the second pointer pReconstructInfo[1] points to the structure with parameters for restructuring Chroms macroblock of plane V. |
The function ippiReconstructChromaInter4x4MB_H264_16s8u_P2R, ippiReconstructChromaInter4x4MB_H264_16s8u_C2R, and ippiReconstructChromaInter4x4_H264High_32s16u_IP2R are declared in the ippvc.h file.
ippiReconstructChromaInter4x4MB_H264_16s8u_C2R is intended for a chrominance combined plane as in the N12 format (UVUVUVUV...).
These functions reconstruct 4x4 inter chroma macroblock (8x8 U macroblock and 8x8 V macroblock) for high profile:
See Example below for the usage of ippiReconstructChromaInter4x4MB_H264_16s8u_P2R.
{
#define D_CBP_CHROMA_DC 0x00001
#define D_CBP_CHROMA_AC 0x1fffe
#define D_CBP_CHROMA_AC_420 0x0001e
#define D_CBP_1ST_CHROMA_DC_BITPOS 17
#define D_CBP_1ST_CHROMA_AC_BITPOS 19
Ipp8u planeU[8*8] =
{ 0x68, 0xaf, 0x16, 0x9a, 0x38, 0x8f, 0x7f, 0x6a,
0x7e, 0x08, 0x01, 0x8c, 0x76, 0x87, 0xa4, 0xc6,
0x9d, 0x47, 0x0f, 0xa1, 0x90, 0x0b, 0x36, 0x47,
0x87, 0x28, 0x0f, 0xb3, 0x24, 0xa4, 0x5c, 0x81,
0x4b, 0x96, 0x33, 0xa3, 0x9f, 0x25, 0x5c, 0x44,
0x61, 0x92, 0x36, 0x9c, 0x07, 0x05, 0x82, 0x9b,
0xc7, 0xb7, 0x99, 0x48, 0x33, 0x9a, 0x75, 0x03,
0xb6, 0x53, 0x9d, 0x03, 0x9c, 0x3a, 0x0b, 0x59 };
Ipp8u planeV[8*8] = {
0x94, 0x22, 0x8a, 0xbc, 0xbb, 0x23, 0x60, 0xbb,
0xba, 0x73, 0x65, 0x14, 0x9a, 0xc6, 0x34, 0xae,
0xaf, 0x16, 0x22, 0x9a, 0x38, 0x6c, 0x54, 0x8f,
0x08, 0x01, 0xb9, 0x8c, 0x76, 0x7b, 0x16, 0x87,
0x47, 0x0f, 0x9f, 0xa1, 0x90, 0x5f, 0x1d, 0x0b,
0x28, 0x0f, 0x4a, 0xb3, 0x24, 0x83, 0x14, 0xa4,
0x96, 0x33, 0x65, 0xa3, 0x9f, 0x3b, 0x08, 0x25,
0x59, 0x2c, 0xb6, 0x30, 0x24, 0x7b, 0x47, 0x6b };
Ipp32s stepUV = 8;
Ipp16s quantTableU[16] = {22,12,34,50,34,33,46,21,18,19,34,28,28,31,15,3};
Ipp16s quantTableV[16] = {32,50,34, 2,46,21,18, 9,34,28,28,31,15,3,12,44};
Ipp16s coef[4+16*4+4+16*4] = {
0, 0, 0, 0,
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, -1, 0,
1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, -1, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-3, -3, -3, -3,
-1, -1, -1, -1, 4, 4, 4, 4, -6, -6, -6, -6, -6, -6, -6, -6,
-6, -6, -6, -6, -6, -6, -6, -6,-64,-64,-64,-64, 7, 1, -4, -3,
7, 1, -4, -3, 2, -2, -2, 2, -1, 5, 10, 11, 11, 15, 15, 9,
-9,-10,-13, 12, -7, -7, -8, 12, -3, -3, 3, 13, -1, -1, 8, 14 };
Ipp32u cbpU = 0x00000014, cbpV = 0x00000014;
Ipp32s QPU = 17, QPV = 31;
Ipp8u bypass_flag = 0;
IppStatus result;
Ipp16s* pCoef;
Ipp32u cbp4x4 = (((cbpU & D_CBP_CHROMA_DC) | ((cbpV & D_CBP_CHROMA_DC) <<
1)) << D_CBP_1ST_CHROMA_DC_BITPOS) | ((cbpU & D_CBP_CHROMA_AC_420) <<
(D_CBP_1ST_CHROMA_AC_BITPOS - 1)) | ((cbpV & D_CBP_CHROMA_AC_420) <<
(D_CBP_1ST_CHROMA_AC_BITPOS + 4 - 1));
pCoef = coef;
result = ippiReconstructChromaInter4x4MB_H264_16s8u_P2R(
&pCoef,
planeU, planeV,
stepUV,
cbp4x4,
QPU,QPV,
quantTableU,quantTableV,
bypass_flag);
}
These functions are used in the H.264 decoder included into Intel IPP Samples. See introduction to H.264.
ippStsNoErr |
Indicates no error. |
ippStsNullPtrErr |
Indicates an error condition if at least one of the specified pointers is NULL. |
ippStsOutOfRangeErr |
ChromaQPU or ChromaQPV is less than 0 or greater than 39. |
Copyright © 2000 - 2010, Intel Corporation. All rights reserved.