Speech intelligibility improvement in noisy reverberant environments based on speech enhancement and inverse filtering

EURASIP Journal on Audio, Speech, and Music Processing

Table 3 Comparison of STI values of different methods

Methods	R1	R2	R3	R4
No equalized RIR	0.99	0.74	0.56	0.40
FIF method	0.99	0.85	0.79	0.71
W-EQ method	1.00	0.88	0.80	0.73
A-EQ method	0.99	0.81	0.74	0.66
Improved FIF method	1.00	0.92	0.84	0.79

R1, R2, R3, and R4 represent the anechoic chamber (T₆₀ < 0.8 s), small classroom (T₆₀ = 0.65 s), large classroom (T₆₀ = 1.39 s), and hall (T₆₀ = 3.57 s), respectively