Motion compensation for predicted 16X8 UV block
SSE2
| FwStatus | fwiMC16x8UV_8u_C1 | ( | const Fw8u * pSrcRef, Fw32s srcStep, const Fw16s * pSrcYData, Fw32s srcYDataStep, Fw8u * pDst, Fw32s dstStep, Fw32s mcType, Fw32s roundControl ); |
| dstStep | Step of the destination array. | |
| mcType | Specifies the type of prediction used for motion compensation. | |
| pDst | Pointer to a location in a destination buffer, a location in a destination array or to a destination result. | |
| pSrcRef | Pointer to a block in the reference plane. | |
| pSrcYData | Pointer to a block of inverse DCT output data. | |
| roundControl | Specifies the type of rounding used for half-pixel approximation. | |
| srcStep | Source buffer step size (width of the buffer in bytes). | |
| srcYDataStep | Source-aligned DCT output data block step size (width of the block in bytes). |
This function adds a 16X8 predicted UV block to a 16X8 decoded residual UV block to reconstruct a source UV block. Prediction is based on the values in a reference UV block, using a compensation type specified by the mcType parameter.
When mcType = FWVC_MC_APX_FF, the reference block is used as the prediction block.
When mcType = FWVC_MC_APX_FH or FWVC_MC_APX_HF, each element of the prediction block is the average of two reference block elements.
When mcType = FWVC_MC_APX_HH, each element of the prediction block is the average of four reference block elements.
The half-sample prediction calculation allows for UV block structure. Neighboring horizontal elements of the U and V blocks have indexes i and i+2 in the UV block.