G711 vs G729
G.711 and G.729 are voice coding methods used for voice encoding in telecommunication networks. Both speech coding methods are standardized in 1990’s, and used in basic applications such as wireless communication, PSTN networks, VoIP (Voice over IP) systems, and switching systems. G.729 is highly compressed compared with G.711. In general, G.711 data rate is 8 times higher than the G.729 data rate. Both methods have evolved during the past decades and have a number of versions as per the ITU-T standard.
G.711 is an ITU-T recommendation for Pulse Code Modulation (PCM) of voice frequencies. G.711 is a commonly used codec in telecommunication channels, which has 64kbps bandwidth. There are two versions of G.711 called μ-law and A-law. A-Law is used in most countries all over the world, while μ-law is primarily used in North America. ITU-T recommendation for G.711 is 8000 samples per second with only a tolerance of + 50 parts per million. Each sample is represented by uniform quantization of 8 bits, which ends up with 64 kbps data rate. G.711 results in very low processing overheads due to the simple algorithms it uses to transform voice signal in to digital format, but leads to poor network efficiency due to inefficient utilization of bandwidth.
There are other variations of G.711 standard such as G.711.0 recommendation, which describes a lossless compression scheme of G.711 bit stream and aimed for transmission over IP services, such as VoIP. Also ITU-T G.711.1 recommendation describes the embedded wideband speech and audio coding algorithm of G.711 standard which operates at higher data rates such as 64, 80 and 96kbps and uses the 16,000 samples per second as default sampling rate.
G.729 is ITU-T recommendation for the coding of speech signals at 8kbps data rate using Conjugate Structure-Algebraic Code Excited Linear Prediction (CS-ACELP). G.729 uses 8000 samples per second while using 16 bit linear PCM as coding method. Data compression delay is 10ms for G.729, also G.729 is optimized to use with actual voice signals which leads to DTMF (Dual Tone Multi-Frequency) tones, and high quality music and fax are not supported reliably using the codec. Therefore, DTMF transmission uses RFC 2833 standard to transmit DTMF digits using RTP payload. Also, the lower bandwidth of 8kbps leads to use the G.729 in Voice Over IP (VoIP) applications easily. Other variants of G.729 are G.729.1, G.729A and G.729B. G.729.1 enables scalable data rates between 8 and 32 kbps. G.729.1 is a wideband speed and audio coding algorithm, which is interoperable with G.729, G.729A and G.729B codecs.
What’s the difference between G711 and G729?
- Both are voice coding systems used in voice communication and standardized by ITU-T.
- Both uses 8000 samples per second for voice signals by applying the Nyquest theory even though G.711 supports 64kbps and G.729 supports 8kbps.
- G.711 concept was introduced in the 1970’s in Bell Systems and standardized in 1988, while G.729 was standardized in 1996.
- G.729 uses special compression algorithms to reduce the data rates, while G.711 requires lowest processing power, when compared with G.729, due to the simple algorithm.
- Both techniques have their own extended versions with small variations.
- Even though G.729 provides low data rates, there are the intellectual property rights that need to be licensed if you need to use G.729,, unlike with G.711.
- Therefore G.711 is supported by most of the devices and interoperability is very simple.
Conversion from one encoding scheme to another will end up with loss of information if there are incompatibilities between codec algorithms. There are systems that measure the quality loss in such scenarios using different indexes such as MOS (Mean Opinion Score) and PSQM (Perceptual Speech Quality Measure).
G.711 and G.729 are voice coding methods specialized to use with telecommunication networks. G.729 operates at 8 times lower data rate compared with G.711 while keeping the similar voice quality using high complex algorithms which leads to higher processing power at the encoding and decoding units.
No related posts.