## 4.1 Definitions, symbols and abbreviations

06.943GPPTSVoice Activity Detector (VAD) for Adaptive Multi Rate (AMR) speech traffic channels

### 4.1.1 Definitions

For the purposes of the present document, the following definitions apply:

**codec: **The combination of an encoder and decoder in series (encoder/decoder).

**compand: **The process of compressing and expanding a signal. In this text, the process is described in terms of PCM [4].

**Decoder**: Generally, a device for the translation of a signal from a digital representation into an analog format. For the present document, a device which converts speech encoded in the format specified in the present document to analog or an equivalent PCM representation.

**DFT**: See Discrete Fourier Transform.

**Discrete Fourier Transform (DFT)**: A method of transforming a time domain sequence into a corresponding frequency domain sequence.

**Encoder**: Generally, a device for the translation of a signal into a digital representation. For the present document, a device which converts speech from an analog or its equivalent PCM representation to the digital representation described in the present document.

**Fast Fourier Transform (FFT)**: An efficient implementation of the Discrete Fourier Transform.

**FFT**: See Fast Fourier Transform.

**Vocoder**: Voice coder.

**frame: **Time interval of 20 ms corresponding to the time segmentation of the speech transcoder.

### 4.1.2 Symbols

For the purposes of the present document, the following symbols apply.

#### 4.1.2.1 Variables

*ch*(*m*)** **channel energy smoothing factor

**(*m*)** **exponential windowing factor

_{}(m)** **estimated spectral deviation between current power spectrum and average long term power spectral estimate

(m)** **spectral peak-to-average ratio

*q*^{(i}^{)}** **quantized channel SNR indices

b(m) burst count

b_{th} burst count threshold

{*d*(*m*)}** **overlapped portion of the frame buffer of input samples

E*ch*(*m,i*)** **channel energy estimate; channel i, subframe m

**E***ch*(*m*) vector of channel energy estimates, 0 i < N_{c}

E_{dB}(m,i)** **estimated log power spectrum

**E***dB*(*m*) vector of log power spectrum estimates, 0 i < N_{c}

average long term power spectral estimate

vector of average long term power spectral estimates, 0 i < N_{c}

E*n*(*m,i*)** **channel noise estimate

**E***n*(*m*) vector of channel noise estimates, 0 i < N_{c}

E_{tn}(m) total estimated noise energy

*Etot*(*m*)** **total channel energy

*E ^{‘}tot*(

*m*)

**modified total channel energy**

*h(m)*** **hysteresis counter

*h _{cnt}*

**hangover count**

*h _{o}*(

*n*)

**overlap-and-add buffer of samples**

*hyster_cnt* hysteresis counter to avoid long term creeping of *update_cnt*

*last_update_cnt*** **previous value of

*update_cnt*

*shp*(*n*)** **sample at the output of the speech encoder high pass filter

*sinewave_flag*** **boolean flag, set TRUE when spectral peak-to-average ratio is greater than 10dB and the spectral deviation is less than DEV_THLD

*SNR*** **Signal to Noise ratio

*SNR _{p}(m)* long-term peak SNR

*SNR _{q}(m)* quantized version of

*SNR*

_{p}(m)*update_cnt*** **counter gating noise estimate update process

*update_flag*** **flag controlling noise estimate updating

*VAD(m)* boolean VAD flag for subframe m

VAD_flag boolean VAD Flag

v(m)** **sum of voice metrics

v_{th} voice metric threshold

#### 4.1.2.2 Constants

* _{H}*

**upper limit for values of**

_{ }**(

*m*)

* _{L}*

**lower limit for values of**

_{ }**(

*m*)

* _{n}*

_{ }channel noise smoothing factor

*p*** **pre-emphasis factor

b_{table}** **table to generate b_{th}

D overlap (delay) in sample intervals

DEV_THLD threshold for setting *sinewave_flag*

E_{floor}** **low threshold for *Etot*(*m*)

*EH*** **high energy endpoint for linear interpolation of

*E*

_{tot}(m)*Einit*** **minimum allowable channel noise initialisation energy

*EL*** **low energy endpoint for linear interpolation of

*E*

_{tot}(m)*E*min** **minimum allowable channel energy

f*H *high channel combining table

f*L *low channel combining table

g(n)** **trapezoidal window, n = 0 to M

G(k)** **frequency domain transformation of g(n)

h_{table}** **table to generate h_{cnt}

HYSTER_CNT_THLD threshold for *hyster_cnt*

L subframe length in samples

M** **DFT sequence length

*Nc*** **number of combined channels

*NOISE_FLOOR_D* low threshold for *Etot*(*m*) in dB

*UPDATE_CNT_THLD*** **threshold for

*update_cnt*

*UPDATE_THLD* threshold for *v(m)*

V** **voice metric table

v_{table}** **table to generate v_{th}

#### 4.1.2.3 Functions

**+** addition

**–** subtraction

***** multiplication

**/** division

** **largest integer x

**AND** Boolean AND

**OR** Boolean OR

### 4.1.3 Abbreviations

**ANSI** American National Standards Institute

**DTX** Discontinuous Transmission

**VAD **Voice Activity Detector

**CAD **Complex Activity Detection

**CNG **Comfort Noise Generation