Intel® Integrated Performance Primitives
Deliberate problems developing high-performance vision, signal, security, and storage applications.

HELP! MPEG4Encode Problem

win847
Beginner
1,086 Views

Hi,

I met link problem in MPEG4Encode codes, i use pca-linux samples.

In mp4emblk.c file, contains ippiCopyBlock_8x8_8u and ippiReconBlock_8x8 functions.But i can't link it with IPP. My download IPP version is 5.0 1.002 for windows.Other IPP funtions can link with.

Please help me.

Best regards.

win847

0 Kudos
5 Replies
win847
Beginner
1,086 Views
Hi, I read document now.Know that ippiCopyBlock_8x8_8u and ippiReconBlock_8x8 contains in IPP4.0 for Linux.But why windows verson don't have these functions?
0 Kudos
win847
Beginner
1,086 Views

Hi,Nikolay

I can't find ippiCopy8x8_8u_C1R and ippiAdd8x8HP_16s8u_C1RS functions either.I download the IPP version is 5.0 for windows.It is used for PXA270 processor.

In ippVc.h

I can find General Motion compensation function as below:

IPPAPI(IppStatus, ippiMCBlock_RoundOff_8u,
(const Ipp8u * pSrc, int srcStep, Ipp8u * pDst, int dstStep,int predictType))
IPPAPI(IppStatus, ippiMCBlock_RoundOn_8u,
(const Ipp8u * pSrc, int srcStep, Ipp8u * pDst, int dstStep,int predictType))
IPPAPI(IppStatus, ippiMCReconBlock_RoundOn,
(const Ipp8u *pSrc, int srcStep, Ipp16s * pSrcResidue, Ipp8u * pDst,int dstStep, int predictType))
IPPAPI(IppStatus, ippiMCReconBlock_RoundOff,
(const Ipp8u *pSrc, int srcStep, Ipp16s * pSrcResidue, Ipp8u * pDst, int dstStep, int predictType))

and find 8 by 8 Inverse Discrete Cosine Transform functions as below:

IPPAPI(IppStatus, ippiDCT8x8Inv_Video_16s_C1, (const Ipp16s * pSrc, Ipp16s * pDst))

IPPAPI(IppStatus, ippiDCT8x8Inv_Video_16s_C1I, (Ipp16s * pSrcDst))

IPPAPI(IppStatus, ippiDCT8x8Inv_Video_16s8u_C1R, (const Ipp16s * pSrc, Ipp8u * pDst, int dstStep))

I think which I need maybe in these functions.Could you help me?

Best Regards

win847

0 Kudos
win847
Beginner
1,086 Views

Hi,Nikolay

Thanks very much.

0 Kudos
win847
Beginner
1,086 Views

Hi,Nikolay

I have another question,I want convert RGB888 color space to YUV422 planar color space.In IPP version 5.0 I use fuction as below.

IPPAPI(IppStatus, ippiBGRToYCbCr422LS_MCU_8u16s_C3P3R, (const Ipp8u * pSrcBGR,
int srcStep,
Ipp16s * pDstMCU[3]))

I want to know how to set srcStep?

Best regards.

win847

0 Kudos
win847
Beginner
1,086 Views

Hi,Nikolay

Thank you very much.

I use ippiRGBToYUV422_8u_C3P3R fuction to convert color space.

My picture size is 320*240,format is RGB888.

and I set srcStep 320*3. and dstStep[3] as below:

dstStep[0]=320,dstStep[1]=160,dstStep[2]=160.

But when I copy YUV data to PXA270 Overlay2 frame buffer, I find that color isn't right.I make U and V data with 0xFF. It shows clear.

Could you help me analyse this problem?

Best Regards.

win847

0 Kudos
Reply