Decoder homing frame

Description

A decoder homing frame is a concept in speech coding, specifically for the Adaptive Multi-Rate Wideband (AMR-WB) codec, as defined in 3GPP specifications such as TS 26.090, TS 26.190, TS 46.020, and TS 46.060. In AMR-WB, there are nine different codec modes, each with a different bit rate. For each mode, a corresponding decoder homing frame exists. These frames are used to synchronise the decoder with the encoder, ensuring that the decoder's internal state matches that of the encoder after a period of transmission or when the encoder sends specific homing frames.

The decoder homing frame is defined as the natural response of the speech encoder to the second and any further encoder homing frame if at least two encoder homing frames were input to the encoder consecutively. In simpler terms, when the encoder receives a sequence of homing frames (special frames that reset the encoder state), it produces a specific output frame that the decoder can recognise. The decoder, upon receiving this frame, can reset its state to match the encoder, thus maintaining synchronisation.

The parameter values for each decoder homing frame are specified in the relevant 3GPP specifications. These parameters include the encoded bits that represent the frame, which are used by the decoder to update its internal state. The decoder homing frames are essential for ensuring that the decoder correctly interprets the encoded speech after a reset or when switching between codec modes. They are particularly important in applications such as Voice over IP (VoIP) and circuit-switched telephony where AMR-WB is used.

The concept of homing frames is not unique to AMR-WB; it is a general technique in speech coding to maintain synchronisation between encoder and decoder. However, the specific definition and parameters for AMR-WB are standardised in 3GPP to ensure interoperability across different implementations. The decoder homing frames are used in conjunction with encoder homing frames, which are sent by the encoder to indicate a reset. When the encoder sends two or more homing frames consecutively, the decoder can expect a specific decoder homing frame as the response, allowing it to reset its state.

In summary, decoder homing frames are a crucial part of the AMR-WB codec, enabling robust synchronisation between encoder and decoder. They are defined for each of the nine codec modes and are used to reset the decoder state when necessary, ensuring high-quality speech transmission.

Purpose & Motivation

Decoder homing frames exist to solve the problem of maintaining synchronisation between the speech encoder and decoder in AMR-WB. In speech coding, the encoder and decoder maintain internal states that must be aligned to correctly reconstruct the speech signal. If the states diverge, due to packet loss, mode switching, or other disruptions, the decoded speech may become distorted or unintelligible. Homing frames provide a mechanism to reset the decoder state to a known condition, ensuring that it matches the encoder's state.

The concept was introduced with the AMR-WB codec to enhance robustness and interoperability. Before homing frames, resynchronisation relied on other methods, such as sending silence frames or using explicit reset signals, which were not always reliable. Homing frames offer a standardised, codec-specific way to achieve synchronisation without additional signalling overhead.

In practical deployments, decoder homing frames are used in voice services over 3GPP networks, including VoLTE and VoWiFi. They help maintain call quality even when the underlying transport experiences errors or when the codec mode is changed. The standardisation of these frames in 3GPP specifications ensures that different vendors' equipment can interoperate seamlessly, which is essential for global roaming and multi-vendor networks.