Skip to main content

Table 10 Mean objective results for babble noise

From: An adaptive a priori SNR estimator for perceptual speech enhancement

Gain SNR PESQ SNR seg HD KurtR
   DD MDD Prop DD MDD Prop DD MDD Prop DD MDD Prop
WF 0 1.971 1.978 1 . 9 9 8 −0.775 −0.693 0 . 6 0 4 0 . 3 9 2 0.401 0.400 1.264 1 . 1 4 8 1.152
  5 2.289 2.316 2 . 3 2 8 1.537 1.665 1 . 7 2 3 0 . 3 5 4 0.396 0.369 1.531 1 . 2 7 9 1.287
  10 2.603 2.632 2 . 6 9 4 4.229 4.411 4 . 5 2 3 0 . 3 0 3 0.323 0.321 1.902 1 . 5 3 4 1.554
LSA 0 1.954 2.009 2 . 0 1 1 −1.007 −0.604 0 . 5 7 0 0 . 3 1 9 0.362 0.360 1.481 1 . 3 4 4 1.347
  5 2.253 2.344 2 . 3 9 9 1.251 1.801 1 . 8 6 4 0 . 2 7 1 0.328 0.309 1.715 1 . 4 5 0 1.456
  10 2.572 2.669 2 . 6 9 6 4.032 4.658 4 . 7 4 3 0 . 2 1 7 0.282 0.208 1.936 1 . 6 3 4 1.648
  1. Bold values denote the best performance