4 General

26.4503GPPCodec for Enhanced Voice Services (EVS)Discontinuous Transmission (DTX)Release 17TS

4.1 Background

The present document specifies the system level aspects of the Discontinuous Transmission (DTX) function of the EVS codec. Specific codec functions that implement the DTX and other EVS functionalities are described in TS 26.445 [2], TS 26.446 [6], TS 26.447 [7], TS 26.448 [8], TS 26.449 [9] and TS 26.451 [11]. The corresponding fixed-point reference C code is specified in TS 26.442 [3] and TS 26.452 [15] using two different sets of basic operators. The corresponding floating-point reference C code is specified in TS 26.443 [4]. The EVS test sequences are given in TS 26.444 [5] and the EVS codec is characterized in TR 26.952 [14].

Discontinuous transmission (DTX) is a mechanism for the EVS Codec to encode the input signal at a lower average rate by taking speech inactivity into account. The DTX scheme is useful for the following purposes:

– to conserve UE power,

– to improve network capacity.

DTX in the transmitting path (uplink) shall be in operation in UEs if commanded so by the network. The UE shall handle DTX in the receiving path (downlink) at any time, regardless of whether DTX in the transmitting path is commanded or not.

4.2 General organization

The default DTX mechanism described in the present document requires the following functions:

– a Signal Activity Detector (SAD) on the transmit (TX) side;

– evaluation of the background acoustic noise on the transmit (TX) side in order to transmit characteristic parameters to the receive (RX) side;

– generation, on the receive (RX) side, of a similar noise called comfort noise during periods where the transmission is switched off.

The Signal Activity Detector (SAD) is defined in TS 26.451 [11] and the comfort noise functions in TS 26.449 [9]. Both are based partly on the EVS Codec and its internal variables defined in TS 26.445 [2].

In addition to these functions, if the parameters arriving at the RX side are detected to be seriously corrupted by errors, the speech or comfort noise shall be generated from substituted data in order to avoid seriously annoying effects for the listener. These functions are defined in TS 26.447 [7].

An overall description of the speech processing parts can be found in TS 26.441 [10]. An overview of the DTX operation is shown in Figure 1.

Figure 1: Block diagram of DTX Operation