Codec Mode Request

Description

The Codec Mode Request (CMR) is a critical control parameter embedded within the speech frames of Adaptive Multi-Rate (AMR), AMR-Wideband (AMR-WB), and Enhanced Voice Services (EVS) codecs in 3GPP systems. It operates within the voice bearer path, typically between the User Equipment (UE) and the core network's Media Gateway (MGW) or Media Resource Function (MRF). The CMR field is part of the codec's payload structure, specified in the AMR and EVS frame formats, and is transmitted in-band alongside compressed speech data. Its primary function is to signal a desired change in the active codec mode—essentially the bitrate and coding scheme—for subsequent speech frames from the receiving end to the transmitting end.

Architecturally, CMR facilitates a closed-loop adaptation system. During a voice call, the receiver (e.g., the far-end UE or network node) continuously assesses channel conditions, such as Block Error Rate (BLER) or signal quality. Based on this assessment, it determines the most appropriate codec mode that balances speech quality and robustness. For instance, in poor radio conditions, a lower bitrate, more error-resilient mode (like AMR 4.75 kbps) might be preferred. The receiver then inserts this requested mode into the CMR field of the next outgoing speech frame it transmits back to the original sender. Upon receiving a frame containing a CMR, the transmitter should adapt its speech encoding to the requested mode for future frames, thereby aligning the transmission with the perceived network conditions at the receiver's side.

Key components involved in CMR operation include the speech codec itself (with its defined modes), the in-band signaling channel within the RTP payload (as per IETF RFCs 4867, 3267, and 3551 adopted by 3GPP), and the rate adaptation algorithms in the UE and network elements. The process is governed by 3GPP specifications detailing codec operation (TS 26. series) and bearer management (TS 29. series). CMR does not function in isolation; it works in tandem with out-of-band Radio Access Bearer (RAB) reconfiguration procedures managed by the Radio Resource Control (RRC) layer for more drastic changes. However, CMR provides a faster, in-session method for fine-grained adaptation.

Its role in the network is pivotal for maintaining consistent voice quality and service reliability. By enabling dynamic bitrate switching, CMR helps conserve radio resources during good conditions (using higher quality modes) and ensures call continuity during degradation (by switching to more robust modes). This directly impacts metrics like Mean Opinion Score (MOS) and call drop rate. In VoLTE and VoNR deployments, CMR is a fundamental part of the voice service enabler, ensuring that the IP-based voice quality adapts efficiently to the underlying LTE or NR radio link.

Purpose & Motivation

CMR was created to solve the fundamental challenge of delivering consistent, high-quality voice service over wireless channels characterized by highly variable and unpredictable conditions, such as fading, interference, and changing user mobility. Prior to adaptive codecs like AMR, fixed-rate codecs (e.g., Full Rate or Enhanced Full Rate in GSM) were used. These codecs operated at a single bitrate, making them suboptimal: they either wasted capacity when channel conditions were good (by not using available bandwidth for higher quality) or suffered from unacceptable quality degradation when conditions were poor, as they lacked a mechanism to trade bitrate for increased error protection dynamically.

The introduction of AMR in 3GPP Release 98 (GSM) and its integration into UMTS (Release 99) marked a paradigm shift, with CMR as its core control mechanism. The purpose of CMR is to enable this adaptation in real-time during an active call. It addresses the limitations of static codecs by providing a low-latency, in-band signaling channel for the receiver to inform the transmitter about the optimal operating point. This was motivated by the need to improve spectrum efficiency and user experience simultaneously. CMR allows the network to maximize voice capacity (by packing more users when using lower rates) while ensuring individual call quality is maintained during signal fluctuations.

Furthermore, as services evolved to wideband (AMR-WB) and super-wideband (EVS) voice, the range of possible codec modes expanded significantly, increasing the potential gain from dynamic adaptation. CMR extended to these codecs, solving the problem of efficiently managing a much larger set of bitrates and bandwidths over IP-based transports like VoLTE. It ensures that the enhanced voice quality of EVS, for example, does not come at the cost of reliability; the codec can seamlessly fall back to a narrower bandwidth or more robust mode when needed, all directed by the CMR mechanism.

Key Features

• In-band signaling within speech frame payload for minimal latency
• Dynamic request for codec mode change (bitrate/bandwidth) from receiver to transmitter
• Supports multiple codec families: AMR, AMR-WB, and EVS
• Enables trade-off between speech quality and error resilience based on radio conditions
• Operates independently but complementary to network-controlled RAB reconfiguration
• Standardized frame format and CMR field location in 3GPP TS 26.103 and related specs

Evolution Across Releases

Defining Specifications