Earlier this year (January 2002) the ITU announced that it had approved a new standard for high-quality digital wideband speech encoding designed to bring significant improvements in terms of interoperability, easier implementation, and improved quality, for wideband voice applications and services across a wide range of communication systems and platforms. The standard, known as recommendation G.722.2 is also referred to as the Adaptive Multi-Rate Wideband (AMR-WB) codec.
Montreal, Canada-based VoiceAge Corporation has prepared a comprehensive white paper entitled Wideband Speech Coding Standards and Applications, which is available as a download from their Web site
ï¿½ Greg Galitzine
The following paragraph is part of a technical contribution submitted to the TIA Engineering Committee working on IP phone specifications.
Technical and qualitative advantages
- ï¿½subjective tests made in several languages (English, Finnish, French, German, Spanish, and Japanese) and performed by several laboratories have shown very good results for the AMR-WB algorithm with data rate 12.65 kbits/s and upwards. For these the quality is equal or better than for ITU-T G.722 algorithm at 64 kbits/s rate.
- The 8.85 kbits/s mode gives quality equal to ITU-T G.722 algorithm at 48 kbits/s rateï¿½
- ï¿½the AMR-WB vocoder has a built-in silence suppression mechanism including a VAD function in the coderï¿½
- ï¿½a CNG function in the decoder, in order to decrease the transmitted data rate during silence phase of the communication and consequently to improve the bandwidth efficiency. This mechanism is quite similar to those used in other speech compression algorithms (e.g. ITU-T G.723.1 or G.729).
- ï¿½AMR-WBï¿½the highest data rate (23.85 kbits/s) has a similar quality on music signal to that of ITU-T G.722 algorithm at 48 kbits/s data rate.
- ï¿½ AMR-WB offers significant lower data rate than the current L16-256 (PCM) mandatory codec chosen by TIA TR-41 and consequently, AMR-WB vocoder can be used for VoIP communication where the LAN might have some traffic overload!
The wireless industry (through 3GPP) selected AMR-WB in 2001 as the mandatory codec for wideband telephony. The ITU-T adopted AMR-WB as Recommendation G.722.2, paving the way towards harmonized wideband telephony standard of wireline and wireless worlds with a fully aligned set of specifications shared by 3GPP and the ITU-T. AMR-WB is a major step in the evolution of wideband speech codecs.
TABLE 1. A comparison of G.722, G.722.1, and G.722.2
||48, 56, 64 (embedded)
||23.85, 23.05, 19.85, 18.25,
15.85, 14.25, 12.65, 8.85, 6.6
||Algebraic Code Excited Linear Prediction (ACELP)
(at 64 kbit/s)
|Poor speech performance in some operating conditions; scope of standard limited to hands-free and low packet loss rates. Good music performance
||Good speech performance at rates 12.65 kbit/s and higher. 15.85 > G.722 @ 56
23.05 > G.722 @ 64
|< 15 MIPS
Exists in Annex B
|ISDN; Video Conferencing
||Same + VoPN
||3G wireless; + Same as G.722.1
AMR-WB, or ITU-T G722.2, is an enabler for the emergence of wideband technology in VoIP. Wideband speech technology provides near face-to-face quality:
- More low-frequency content compared to narrowband telephony translates into increased naturalness, presence, and comfort.
- More high-frequency content compared to narrowband telephony translates into greater intelligibility.
AMR-WB is already present in video and audio conferencing systems. It will be introduced through high-tier equipments such as PCs, PDAs, wireless multimedia handsets, and IP phones. Enterprises are in a position to immediately take advantage of the wideband technology for IP-to-IP communications. The global market will follow once they start using digital endpoints: IP phones, wireless phones, PDAs, and PCs.
Nowadays standards include much functionality as specified in the terms of references as part of the standardization process. It is a new trend to extend those functionalities while respecting the
"interoperability" of the standard, i.e., without affecting the compressed bit stream. VoiceAge has introduced
"extensions" as functions available to the application making use of the codec.
VoiceAge's extensions are split between "bandwidth extensions" enabling a higher audio bandwidth with the same core technology and robustness to packet network impairments to limit the number of packets rejected and also allow a
"smaller" jitter buffer resulting in less end-to-end delay for conversations.
For more information, or to download the white paper Wideband Speech Coding Standards and Applications, please visit the company online at
To The October 2002 Table Of Contents ]