AMRNB

AMRNB, short for Adaptive Multi-Rate Narrowband, is a speech codec developed for mobile networks. As part of the AMR family defined by 3GPP, AMRNB encodes voice in a narrowband channel suitable for traditional telephony and early GSM/UMTS networks.

Technical characteristics: It operates at eight bitrates: 4.75, 5.15, 5.9, 6.7, 7.4, 7.95, 10.2, and 12.2 kbps, using 20 ms frames. The codec is based on algebraic CELP (ACELP) with rate switching to adapt to network conditions. It includes features to improve robustness and efficiency, such as voice activity detection, discontinuous transmission, and comfort noise generation, as well as error concealment for packet loss.

Usage and deployment: AMRNB became a standard for speech in 2G and early 3G networks. It has been widely implemented in base stations, mobile devices, and VoIP gateways, enabling efficient voice calls over GSM, GPRS, and early UMTS. It is typically used where backward compatibility with legacy networks is required. AMRNB is complemented by AMR-WB, which provides higher quality at wider bandwidth, and together they form the AMR codec family standardized by 3GPP.

Current status: While newer codecs such as AMR-WB+ and EVS have superseded AMRNB in new deployments, AMRNB remains present in many legacy systems and devices. The AMR family remains defined by 3GPP and is supported by multiple implementations, including open-source decoders and commercial codecs, with licensing typically governed by patent and 3GPP terms.