What is G.729?

G.729 is a voice audio data compression algorithm that compresses voice audio in chunks of 10 milliseconds. Music or tones such as Dual-tone Multi-frequency (DTMF) signaling or fax tones cannot be transported reliably with this codec, and thus use G.711 or out-of-band methods to transport these signals.

G.729 is a standard for Internet Protocol (IP) Private Branch Exchange (PBX) vendors, as well as for the public switched telephone network (PSTN). G.729 digitizes analog voice signals producing output at 8 kilobits per second (Kbps) with 8:1 compression and employs an algorithm called conjugate-structure algebraic code-excited linear prediction (CS-ACELP).

G.729 has several extensions or annexes. The two most common are G.729A and G.729AB. In G.729A, input frames are 10 milliseconds (10 ms) in duration and generated frames contain 80 bits. The input and output contain 16-bit pulse-code modulation (PCM) samples converted from or to 8-Kbps compressed data. The total algorithmic delay is 15 ms.

In G.729AB, the parameters are the same as in G.729A with the addition of voice activation detection (VAD) in which the effective signal bandwidth is reduced when there is no audio input. A technology known as comfort noise generation (CNG) produces a small amount of background pink noise during pauses in speech to avoid user distraction that can be caused by intervals of absolute silence.