MPEG-4 Encoder

[sjoe1]

Hi,

I am trying to optimize MPEG-4 encoder for WinCE using IPP PCA library functions. I had a few doubts regarding some of the functions given, any help would be greatly appreciated.

ippiTransRecBlockCeof_intra_MPEG4 requires a quantization parameter buffer. I am assuming that it refers to the quantization scales of the blocks surrounding the current block. But what is the order in which the quantization parameters of the blocks(topmost left, topmost right,on the left)be given?

I tried encoding the AC coefficients using ppiEncodeVLCZigzag_IntraACVLC_MPEG4_16s1u but it is generating erroneous output. Does the "pattern" input given to function have values only 1 and 0 indicating whether theblock is coded or not. Also where should the pointer to the quantized coefficients begin from. (DC value or the value next to it). I had tried all combination but cannot find the problem, could anyone give me some pointer regarding this?

Thanks,

Sjoe

[nlj]

Hi,

I am also facing a problem in using IPP call for VLC coefficient encoding in MPEG-4 encoder. I am using IPP library ver 4.0 for PCA processor, on Windows.

When I make a call to ippiEncodeVLCZigzag_IntraACVLC_MPEG4_16s1u

in my code, to the output buffer, 4 bytes are written always. (it should be variable length data - depending on how 63 ac coeffs are getting encoded - right ?)

In my code, some other IPP calls- the ones forDCT, IDCT are working fine.

But for the above IPP, getting erroneous results..

Please help !

thanks & regards,

nlj

[Vladimir_Dudnik]

Hi,

could you please submit your issue through premier.intel.com?

Regards,

Vladimir