3 Definitions, abbreviations and mathematical expressions
26.4453GPPCodec for Enhanced Voice Services (EVS)Detailed algorithmic descriptionRelease 15TS
3.1 Definitions
For the purposes of the present document, the terms and definitions given in TR 21.905 [1] and the following apply. A term defined in the present document takes precedence over the definition of the same term, if any, in TR 21.905 [1].
frame: an array of audio samples spanning 20ms time duration.
EVS codec: The EVS codec includes two operational modes: EVS Primary operational mode (EVS Primary mode) and EVS AMR-WB Inter-Operable modes (EVS AMR-WB IO mode). When using the EVS AMR-WB IO mode the speech frames are bitstream interoperable with the AMR-WB codec [9]. Frames generated by an EVS AMR-WB IO mode encoder can be decoded by an AMR-WB decoder, without the need for transcoding. Likewise, frames generated by an AMR-WB encoder can be decoded by an EVS AMR-WB IO mode decoder, without the need for transcoding.
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 SID (3GPP TS 26.441 [2]). The EVS Primary mode can encode narrowband, wideband, super-wideband and fullband signals. None of these bit-rates are interoperable with the AMR-WB codec.
EVS AMR-WB IO mode: Includes 9 codec modes and SID. All are bitstream interoperable with the AMR-WB codec (3GPP TS 26.171 [37]).
Operational mode: Used for the EVS codec to distinguish between EVS Primary mode and EVS AMR-WB IO mode.
3.2 Abbreviations
For the purposes of the present document, the abbreviations given in TR 21.905 [1] and the following apply. An abbreviation defined in the present document takes precedence over the definition of the same abbreviation, if any, in TR 21.905 [1].
ACELP Algebraic Code-Excited Linear Prediction
AMR Adaptive Multi Rate (codec)
AMR-NB Adaptive Multi Rate Narrowband (codec) – Also referred to as AMR
AMR-WB Adaptive Multi Rate Wideband (codec)
AR Auto-Regressive
AVQ Algebraic Vector Quantization
CELP Code-Excited Linear Prediction
CLDFB Complex Low Delay Filter Bank
CMR Codec Mode Request
CNG Comfort Noise Generator
DCT Discrete Cosine Transformation
DFT Discrete Fourier Transform
DTX Discontinuous Transmission
EFR Enhanced Full Rate (codec)
EVS Enhanced Voice Services
FB Fullband
FCB Fixed Codebook
FEC Frame Erasure Concealment
FFT Fast Fourier Transform
FIR Finite Impulse Response
FT Frame Type
GC Generic Coding
HF High Frequency
HP High Pass
IIR Infinite Impulse Response
IP Internet Protocol
ISF Immittance Spectral Frequency
ISP Immittance Spectral Pair
ISPP Interleaved Single-Pulse Permutation
JBM Jitter Buffer Management
LD Low Delay
LP Linear Prediction
LPF Low Pass Filter
LSB Least Significant Bit
LSF Line Spectral Frequency
LSP Line Spectral Pair
LTP Long-Term Prediction
MA Moving Average
MDCT Modified Discrete Cosine Transform
MRLVQ Multi-Rate Lattice Vector Quantization
MSB Most Significant Bit
MSVQ Multi-Stage Vector Quantization
MTSI Multimedia Telephony Service for IMS
NB Narrowband
OL Open-Loop
PCPRS Path-Choose Pulse Replacement Search
PS Packet Switched
PSTN Public Switched Telephone Network
QMF Quadrature Mirror Filter (See also CLDFB)
SAD Signal Activity Detection
SDP Session Description Protocol
SID Silence Insertion Descriptor
SNR Signal-to-Noise Ratio
SWB Super Wideband
TC Transition Coding
ToC Table of Contents
UC Unvoiced Coding
VC Voiced Coding
VMR-WB Variable Rate Multimode Wideband
VQ Vector Quantization
WB Wideband
WMOPS Weighted Millions of Operations Per Second
3.3 Mathematical Expressions
For the purposes of the present document, the following conventions apply to mathematical expressions:
indicates the largest integer less than or equal to x: 
, 
and 
;
indicates the smallest integer greater than or equal to x: 
, 
and 
|x| indicates the absolute value of x: |17| = 17, |–17| = 17;
min(x0,…xN–1) indicates the minimum of x0,…, xN–1, N being the number of components;
max(x0,…xN–1) indicates the maximum of x0, …, xN–1;
AT indicates the transpose of matrix A;
indicates the remainder after dividing x by y: 
;
round(x) is traditional rounding: round(x) = 
;
exp(x) is equivalent to ex where e is the base of the natural algorithm;
indicates summation;
indicates product;
Unless otherwise specified, log(x) denotes logarithm at base 10 throughout this Recommendation.