Hi,
can I use the same pointer as arg #1 and arg #4 when calling ippiDequantTransformResidualAndAdd_H264_16s_C1I?
That is:
ippiDequantTransformResidualAndAdd_H264_16s_C1I(
pPredDst,
pSrcDst,
pDC,
pPredDst, // same pointer
PredDstStep,
PredDstStep, // same step
QP,
AC);
The purpose is when the transformed-quantified predictor is already stored in the frame that is to be reconstructed.
Other question: does there exist a version of ippiPredictIntra_4x4_H264_8u_C1IR taking an explicit pDst? That is:
ippiPredictIntra_4x4_H264_8u_C1IR(
Ipp8u const* pSrc,
Ipp32s srcStep,
Ipp8u* pDst,
Ipp32s dstStep,
IppIntra4x4PredMode predMode,
Ipp32s availability);
instead of the existing:
ippiPredictIntra_4x4_H264_8u_C1IR(
Ipp8u* pSrcDst,
Ipp32s srcdstStep,
IppIntra4x4PredMode predMode,
Ipp32s availability);
This would solve the first issue for me.
(same question for 16x16).
Cheers,
Jerome.
