Enhanced Voice Services Primary Mode

Description

The EVS Primary mode is the core operating mode of the Enhanced Voice Services (EVS) codec, introduced in 3GPP Release 12. It represents the native, non-interoperable mode that delivers the full capabilities of EVS, including superior voice quality, enhanced efficiency, and support for a wide range of audio bandwidths. The EVS Primary mode is specified in 3GPP TS 26.441 and TS 26.445, and its usage in IMS telephony is described in TS 26.114. Unlike the EVS AMR-WB IO mode, which is designed for backward compatibility with AMR-WB, the Primary mode is optimized for EVS-only communication, where both endpoints support the EVS codec.

The EVS Primary mode includes 11 bit-rates for fixed-rate or multi-rate operation, 1 average bit-rate for variable bit-rate operation, and 1 bit-rate for Silence Insertion Descriptor (SID) frames. These bit-rates range from 5.9 kbps to 128 kbps, covering narrowband (NB), wideband (WB), super-wideband (SWB), and fullband (FB) audio. The codec can encode signals with sampling rates up to 48 kHz, providing exceptional voice and music quality. The Primary mode employs advanced coding techniques such as Algebraic Code Excited Linear Prediction (ACELP) for lower bit-rates and Modified Discrete Cosine Transform (MDCT) for higher bit-rates, ensuring high quality across the entire range. The variable bit-rate operation allows the codec to adapt its rate based on network conditions and signal characteristics, optimizing quality and capacity.

The EVS Primary mode is not bitstream interoperable with AMR-WB. This means that an AMR-WB decoder cannot decode EVS Primary bitstreams, and vice versa. Therefore, when an EVS call is established, both endpoints must support EVS Primary mode to use it. If one endpoint only supports AMR-WB, the call will use the EVS AMR-WB IO mode or fall back to AMR-WB. The Primary mode is typically negotiated using SDP offer/answer with the EVS codec and its modes. The EVS codec can switch between Primary and AMR-WB IO modes during a call, depending on the far-end capabilities and network conditions, ensuring seamless interworking.

The EVS Primary mode supports Discontinuous Transmission (DTX) and comfort noise generation using SID frames, which reduces bandwidth during silence periods. It also includes robust error concealment mechanisms to handle packet loss, ensuring acceptable voice quality in adverse network conditions. The codec is designed to be robust to jitter and delay, making it suitable for VoIP and VoLTE services. The EVS Primary mode also supports channel-aware mode, which adapts the codec rate based on radio link conditions, further enhancing quality and capacity.

In summary, the EVS Primary mode is the advanced, native mode of the EVS codec, providing the highest voice quality and efficiency for EVS-capable devices. It is a key enabler for enhanced voice services in LTE and 5G networks, offering a significant improvement over AMR-WB. Its non-interoperability with AMR-WB is a deliberate design choice to allow the use of advanced coding tools without the constraints of backward compatibility, while the AMR-WB IO mode ensures interworking when needed.

Purpose & Motivation

The EVS Primary mode was created to provide a state-of-the-art voice codec that significantly outperforms AMR-WB in terms of voice quality, capacity, and flexibility. By Release 12, mobile operators were seeking to improve voice services in LTE networks, which were primarily using AMR-WB. While AMR-WB offered good wideband quality, it was limited in its ability to deliver super-wideband and fullband audio, and its efficiency was not optimal for the growing demand for high-definition voice and multimedia telephony. The EVS Primary mode was designed to address these limitations by introducing a new codec with a wide range of bit-rates and audio bandwidths, enabling operators to deliver superior voice quality and to optimize network capacity.

The Primary mode was motivated by the need for a codec that could support both voice and music, with high quality for applications such as HD voice, immersive telephony, and multimedia sharing. It also needed to be efficient enough to increase the number of voice calls per cell, reducing the cost per call. The variable bit-rate operation allows the codec to adapt to network conditions, maintaining quality while conserving bandwidth. The inclusion of fullband audio (up to 48 kHz) enables a new level of realism in voice communication, often referred to as 'full-HD voice'. The Primary mode also supports advanced features like DTX and error concealment, which are essential for VoIP over LTE.

The decision to make the Primary mode non-interoperable with AMR-WB was deliberate: it allowed the codec designers to use advanced coding tools without the constraints of backward compatibility, resulting in better performance. However, to ensure that EVS devices could still communicate with AMR-WB devices, the EVS AMR-WB IO mode was included as a separate mode. This dual-mode approach enables a smooth transition to EVS while maintaining interoperability with legacy networks. The EVS Primary mode thus represents a significant leap forward in voice coding, and it has been adopted in 5G networks as well, where it continues to provide high-quality voice services.