5.3.1 General description

26.4453GPPCodec for Enhanced Voice Services (EVS)Detailed algorithmic descriptionRelease 15TS

The decision algorithm for selecting the MDCT Coding mode, instead of CELP, is described in subclause 5.1.16. In general, the MDCT coding mode is selected whenever a signal is not speech or the bit rate for encoding is high enough for the transform coder. It always operates on full input signal independent of sample rate and selected audio bandwidth.

The EVS codec supports two major MDCT coding modes, the MDCT based TCX (TCX) and the High Quality MDCT coder (HQ). Both operate with different configurations depending on the bit rate for encoding. For some operating points, the MDCT coders are switched for optimized efficiency. The two MDCT selector algorithms are described in subclauses 5.1.14.2 and 5.1.14.3. The following table describes the used mode configurations depending on bit rate and bandwidth.

Table 76: Overview of supported MDCT coding modes

Bitrate (kbps)	Bandwidth	MDCT Classifier	MDCT mode
7.2, 8	NB		HQ low-rate
9.6	NB, WB,SWB		TCX low-rate
13.2	NB, WB, SWB	MDCT Selector	Switching HQ low-rate and TCX mid-rate
16.4	NB, WB, SWB FB
24.4	NB, WB		TCX mid-rate
32	WB		HQ high-rate
24.4	SWB,FB	MDCT Selector II	Switching HQ high-rate and TCX mid-rate
32	SWB,FB
48	WB,SWB,FB		TCX high-rate
64	WB,SWB,FB		HQ high-rate
96,128	WB,SWB,FB		TCX high-rate

The next subclauses describe the common time-to-frequency transformation for both major MDCT modes. In the following the TCX and the HQ coder are described including their modules and configurations.