Performance of adaptive multirate (AMR) voice in GSM and WCDMA

This paper studies the performance of adaptive multirate (AMR) voice service in GSM and WCDMA radio networks. The AMR voice codec and its utilization in GSM and WCDMA radio networks are first introduced. The effect of AMR to GSM and WCDMA spectral efficiency is evaluated. It is shown that AMR can improve GSM spectral efficiency 100-150% compared to the enhanced full rate (EFR) codec. AMR further allows the utilization of half-rate codec improving the hardware efficiency in GSM. In WCDMA networks, the AMR allows to make a trade-off between voice capacity and AMR bit rate. The paper also presents the future evolution of AMR voice coding for further enhancements of spectral efficiency and voice quality. The source based rate adaptation is shown to improve the AMR spectral efficiency up to 25%. The wideband AMR will bring improved audio quality for the wireless and fixed network.