Table 6. WRR Performance of benchmark MCSE and the proposed MCSE under non-stationary noises of the AURORA dataset.

Database	Noise	SNR/dB	BAV-MCSE	DWT-CNN-MCSE	DIFFERENCES
			WRR (%)	WRR (%)	WRR (%)
AURORA (Non-Stationary Noises)	Airport	−10	5.82	70.55	64.73
		−5	12.32	72.51	60.19
		0	19.06	78.75	59.69
		5	36.14	77.44	41.3
		10	67.26	67.15	−0.11
		15	88.88	75.44	−13.44
	Babble	−10	4.04	68.50	64.46
		−5	7.12	70.25	63.13
		0	17.56	70.32	52.76
		5	35	66.44	31.44
		10	74.18	73.79	−0.39
		15	90.64	61.49	−29.15
	Car	−10	7.26	72.50	65.24
		−5	13.26	74.60	61.34
		0	16.55	80.49	63.94
		5	35.16	77.44	42.28
		10	67.2	81.49	14.29
		15	92.02	78.8	−13.22
	Exhibition	−10	6.54	65.15	58.61
		−5	11.54	68.25	56.71
		0	20.23	80.45	60.22
		5	44.66	77.73	33.07
		10	77.72	76.42	−1.3
		15	91.46	73.75	−17.71
	Restaurant	−10	4.54	72.50	67.96
		−5	6.38	73.50	67.12
		0	12.12	80.39	68.27
		5	38.24	77.75	39.51
		10	55.56	76.35	20.79
		15	75.2	74.48	−0.72