Adaptive Multi-Rate Full Rate Speech

Description

AFS, or Adaptive Multi-Rate Full Rate Speech, is a specific operational mode of the Adaptive Multi-Rate (AMR) speech codec standardized by 3GPP for GSM/EDGE Radio Access Network (GERAN) systems. It functions on the full-rate traffic channel (TCH/F), which has a gross bit rate of 22.8 kbit/s. The core principle of AFS is its adaptive nature: it does not operate at a single fixed bit rate. Instead, it selects from a set of eight source codec modes, ranging from 4.75 kbit/s to 12.2 kbit/s (including the ETSI GSM Enhanced Full Rate codec mode). The remaining channel capacity, after allocating bits for the speech payload, is used for forward error correction (FEC) through channel coding. The network, specifically the Base Station Controller (BSC), dynamically chooses the optimal codec mode based on real-time measurements of radio channel quality, such as signal strength and bit error rate.

Architecturally, AFS implementation involves both the Mobile Station (MS) and the network elements. The speech encoding occurs in the MS and the Transcoder Unit (TCU), often located in the Base Station Controller (BSC) or Media Gateway (MGW). The BSC plays a pivotal role as it receives measurement reports from the MS and the Base Transceiver Station (BTS) regarding the current radio link quality. Using this information, the BSC executes the AMR mode adaptation algorithm to decide whether to command a change to a different codec mode—for example, shifting from a high-bit-rate mode (like 12.2 kbit/s) offering best speech quality to a more robust, lower-bit-rate mode (like 4.75 kbit/s) with stronger FEC when interference increases. This command is sent via in-band signaling within the speech frame.

The role of AFS in the network is to maximize spectral efficiency and enhance voice quality robustness. By allowing the speech codec bit rate to be traded for increased error protection, AFS provides a significant gain in tolerance against channel errors compared to older, fixed-rate codecs like the GSM Full Rate (FR) or Enhanced Full Rate (EFR). This translates directly into improved voice quality at cell edges or in poor coverage areas and can increase network capacity by allowing acceptable quality calls at lower carrier-to-interference (C/I) ratios. The adaptation is typically slow, on the order of seconds, to match the relatively slow fading characteristics of the radio channel and to avoid excessive mode switching. AFS is a key component of the AMR codec's overall functionality, which also includes the Adaptive Multi-Rate Half Rate Speech (AHS) mode for operation on half-rate channels, providing further capacity gains.

Purpose & Motivation

AFS was created to address the limitations of fixed-rate speech codecs in GSM networks, primarily the trade-off between speech quality and error robustness. Pre-AMR codecs like GSM Full Rate (FR) and Enhanced Full Rate (EFR) used a static allocation of bits for speech and channel coding. In good radio conditions, this was inefficient as excess channel coding bits were wasted, while in poor conditions, the fixed level of protection was often insufficient, leading to severe voice degradation or dropped calls. The industry needed a more intelligent codec that could dynamically optimize this trade-off.

The motivation for AFS and the broader AMR codec was driven by the need for improved voice service quality and increased network capacity without requiring new spectrum or additional cell sites. By introducing adaptability, the same physical radio channel could support higher quality speech when conditions were good and more robust, albeit slightly lower quality, speech when conditions deteriorated. This adaptability directly solved the problem of inconsistent voice quality experienced by mobile users, especially in challenging radio environments. It allowed operators to improve customer satisfaction and potentially tighten frequency reuse patterns, thereby increasing the number of simultaneous calls the network could handle.

Historically, AFS was part of the AMR codec's introduction in 3GPP Release 4 (though often associated with later releases for specific features), marking a significant evolution from the GSM EFR codec. It provided a backward-compatible upgrade path for existing GSM networks, as the full-rate channel structure remained unchanged; only the interpretation of the bitstream was adapted. This made AFS deployment relatively straightforward for operators, enabling them to reap the benefits of adaptive speech coding across their existing infrastructure, paving the way for more efficient use of the valuable radio spectrum dedicated to voice services.

Key Features

• Dynamic codec mode selection from eight rates (4.75, 5.15, 5.90, 6.70, 7.40, 7.95, 10.2, 12.2 kbit/s)
• Operates on the GSM Full-Rate Traffic Channel (TCH/F) at 22.8 kbit/s
• Link adaptation controlled by the BSC based on radio quality measurements
• In-band signaling for codec mode commands and indications
• Enhanced error robustness through adaptive channel coding allocation
• Backward compatibility with GSM channel structure for ease of deployment

Evolution Across Releases

Defining Specifications