Skip to main content

Advertisement

Table 1 SINR-out for different algorithms (dB)

From: Speech enhancement with an acoustic vector sensor: an effective adaptive beamforming and post-filtering approach

Algorithm ZPF[4] NCFMV[5] NCFMV-ZPF[6] NCFMV-EWPF SINR-input (dB)
Trial 1 (navs(t) = 0 and s i (t) ≠ 0) 2.7 12.6 14.8 26.2 0
Trial 2 (navs(t) = 0 and s i (t) ≠ 0) 7.8 12.8 16.4 34.0 5
Trial 3 (navs(t) = 0 and s i (t) ≠ 0) 13.1 13.4 18.3 28.3 10
Trial 4 (navs(t) ≠ 0 and s i (t) = 0) 8.1 2.0 7.8 8.3 0
Trial 5 (navs(t) ≠ 0 and s i (t) = 0) 13.5 6.5 13.5 13.2 5
Trial 6 (navs(t) ≠ 0 and s i (t) = 0) 17.6 9.1 17.0 16.5 10
Trial 7 (navs(t) ≠ 0 and s i (t) ≠ 0) 3.1 8.1 11.9 14.2 0
Trial 8 (navs(t) ≠ 0 and s i (t) ≠ 0) 8.3 10.3 14.7 18.0 5
Trial 9 (navs(t) ≠ 0 and s i (t) ≠ 0) 13.6 12.4 18.9 21.2 10