×

Warning message

  • Cisco Support Forums is in Read Only mode while the site is being migrated.
  • Cisco Support Forums is in Read Only mode while the site is being migrated.

DSP and complexity

Unanswered Question
allan.wells Tue, 09/30/2003 - 13:22
User Badges:

Hi,


The primary difference is the processing power needed for each individual codec g.729 Vs 729a as an example not much difference from a quality standpoint.


The difference being the number of medium VS high complexity VoCoders

calls

processed example

g.729a medium 17 MIPS Vs 34 MIPS g729 high complexity.

Smaller codebook used for g.729a


The G.729 speech codec uses the conjugate-structure algebraic code excited

linear-prediction(CS-ACELP) algorithm, which is an analysis-by-synthesis

algorithm and belongs to the class of speech coding algorithms known as code

excited linear prediction (CELP). Therefore, the speech decoder is primarily

a subset of the speech encoder. For every 10-ms speech frame, thespeech

signal is analyzed to extract the parameters of the CELP model

(linear-prediction filter coefficients, adaptive and fixed-codebook indices

and gains). The G.729A is the annex A forG.729 to reduce complexity of the

CS-ACELP speech codec .




Analog voice is received by a CODEC (coder / de-coder)


* Complex circuitry then samples the voice and compares the voice to a



codebook. This codebook is used by both the coders and decoders.


* Once sampled and matched with the codebook the voice is now


presented as data


* Some codebooks are of different sizes and have a different fidelity


* These codebooks are "recommended standards" that copies voice


samples into different sizes


* Codec standards:


­


G.711 changes voice into data but uses the same amount of space as an

analog


method. Uses 64Kcodebook. This codebook is used by both the coders and

decoders.


* Once sampled and matched with the codebook the voice is now



presented as data


* Some codebooks are of different sizes and have a different fidelity


* These codebooks are "recommended standards" that copies voice


samples into different sizes


* Once sampled and matched with the codebook the voice is now


presented as data


* Some codebooks are of different sizes and have a different fidelity


* These codebooks are "recommended standards" that copies voice


samples into different sizes


* Codec standards:


­


G.711 changes voice into data but uses the same amount of space as an

analog


method. Uses 64K




­ G.729 Uses a large codebook but sounds good at 1/8 of G.711. Uses about 8K


­ G.729A Codebook smaller than G.729






* Complex circuitry then samples the voice and compares the voice to a


codebook. This codebook is used by both the coders and decoders.


* Once sampled and matched with the codebook the voice is now


presented as data


* Some codebooks are of different sizes and have a different fidelity


* These codebooks are "recommended standards" that copies voice


samples into different sizes


* Codec standards:


­


G.711 changes voice into data but uses the same amount of space as an

analog


method. Uses 64K




­ G.729 Uses a large codebook but sounds good at 1/8 of G.711. Uses about 8K


­ G.729A Codebook smaller than G.729



http://www.cisco.com/en/US/tech/tk652/tk701/technologies_tech_note09186a00800b6710.shtml


regards


Allan


Actions

This Discussion