endian for ippsEncode_G726_16s8u?

[gwshieh]

Hi,

As far as I know, there are 2 versions of G726 definitions from ITU and IETF. The one from IETF is little endian and ITU is big endian in terms of the packing of codeword.

Is this function ippsEncode_G726_16s8u has any assumption about this? The sample g726 codec seems to be little endian.

thanks,
guoweis

[Igor_B_Intel1]

Hi,

Function ippsEncode_G726_16s8u produces codewords in separete bytes. For example for32 kbp mode, size of one codeword -4 bits, so only4 bits in each byte are valid. You can pack it as you would like to have. G726 sample inIPP 5.0 is done according RFC 3551.

Igor S. Belyakov