Do you know how to umc_h264_dec_con(sample source) convert the high level function into the low level function?

[souler3]

Hello.

I've been using IPP5.3 H.264 Decoder in my project.

Do you know how to umc_h264_dec_con(sample source) convert the high level function into the low level function? Do you know the best way or the sample code?

I've been reading the ippiman.pdf in the section of H.264 Decoder Functions. I attempt to decode the intra4x4 macroblock.(Figure 16-53)

The follwing has used a high level function in the sample code.
ReconstructLumaIntraMB(
(CoeffsPtr*)&sd->m_pCoeffBlocksRead,
(PlanePtrY)sd->m_pYPlane + offsetY,
rec_pitch_luma,
(IppIntra4x4PredMode_H264 *) pMBIntraTypes,
sd->m_cur_mb.LocalMacroblockInfo->cbp4x4_luma,
lumaQP,
edge_type,
sd->bit_depth_luma);

I want to replace the ReconstructLumaIntraMB(...) by the low level function(ex:ippiPredictIntra_4x4_H264(...), ippiDequantTransformResidualAndAdd_H264(...) or ippiDequantTransformResidual_H264(...)).

But I don't know parameters for the low level function.

I've been looking for advices.

Please help me.

[Vladimir_Dudnik]

Hello,

We've described the meaning of each parameter for each IPP function in manual. And we provide an example of complete H.264 decoder in source code. What exactly you are looking for?

Regards,
Vladimir

[souler3]

Thank you for your reply.

I'm curious about the first parameter of ReconstructLumaIntraMB(...).

Why did you use the bitstream for the reconstruction?

I think that the reconstruction shoulduse the coefficients decoded.

(ex : sd->m_cur_mb.MacroblockCoeffsInfo->numCoeff)

In my understanding, numCoeff are decoded in accordance with the 9.2 of JVTG050.

Why did you use the bitstream instead of the coefficients decoded?

[Vladimir_Dudnik]

Why do you decide this function use bitstream for reconstruction? According function description in IPP documentation it uses pointer to decoded coefficients.

Vladimir