Forums

G.729 codec

Started by ramialhasan May 22, 2005
I wonder if any have any idea about the ITU G.729 sample implementation
sample code. when we use their decoder application, then generated PCM
16-bit file is almost the same size as the input bitstream g.729 file,
and we know that this is a compressed format, is there a problem, or
those samples are assuming something special?

Please help, Thanks
Rami

ramialhasan wrote:
> I wonder if any have any idea about the ITU G.729 sample implementation > sample code. when we use their decoder application, then generated PCM > 16-bit file is almost the same size as the input bitstream g.729 file, > and we know that this is a compressed format, is there a problem, or > those samples are assuming something special?
There is nothing wrong. You just haven't read what is supposed to be in the compressed file, or tried looking (clue: it isn't a compact bitstream). Codecs are not generally defined with a bit packing format, since this is very application dependant. The popularity of G.729 for VoIP means a packing format was defined as an add on to the original spec. Most G.729 for VoIP uses this format. Some early users, such as Cisco, defined their own packing format before it was standardised. You can still find people using these non-standard formats. For something like streaming over wireless, you would generally use something very different from the VoIP packing format. You would intermix redundant bits with the data bits, and apply most of the redundancy to the most important bits (e.g. the MSBs of the energy word). Regards, Steve
Thanks Steve,

Actually I am interested in decoding the g.729 payload of the cisco RTP
voice, and I am trying to use the decoder to convert the payload into
PCM16bit. So what should I do on the payload before passing it to the
decoder, to get the required PCM data, is there any conversion on the
input or output values before passing it to the decoder?

Thanks for your help.