Enhanced Voice Services (EVS) Codec

VOCAL offers the EVS codec software as a standalone algorithm, as part of a modular software library suite, as well as with a VoIP stack. VOCAL’s software is written in ANSI C and does not make use of any GPL licensed software components.

Super-wideband Audio Coding with Backward Compatibility for Narrowband/Wideband Operation
Mobile, VoIP and Voice Conferencing for LTE
Optimized for DSPs, RiSC, CISC processors
3GPP TS 26.441-451 compliant

EVS Overview

Enhanced Voice Services (EVS), standardized by 3GPP, is a multi-rate audio codec supporting an audio bandwidth up to 20kHz with supported use for conversational speech and music. This codec provides enhanced voice quality, better compression rates, and network features over previously standardized 3GPP codecs. EVS codec software was designed for integration with LTE and packet-switch mobile communication networks.

In order to support efficient compression of both speech and music input audio signals, EVS has separate compression modes. Algebraic Code-Excitation Linear Prediction (ACELP) is used for speech signals and the Modified Discrete Cosine Transform (MDCT). For backward compatibility with previous 3GPP standards for narrowband and wideband voice, AMR-WB Inter-Operable mode is implemented. The EVS software also offers a channel-aware mode for additional robustness to channel errors. This is achieved by combining the jitter buffer management (JBM) with partial redundancy. The redundant information is used to synthesize a lost frame or packet.