Skip to main content

Table 9 Mean objective results for pink 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 1.960 1.944 2 . 0 0 4 −0.202 −0.138 0 . 3 2 6 0.441 0.451 0 . 4 0 1 1.064 1 . 0 0 9 1.028
  5 2.401 2.412 2 . 4 6 2 2.530 2.662 3 . 1 4 8 0.399 0.395 0 . 3 5 8 1.253 1 . 0 7 0 1.103
  10 2.748 2.776 2 . 7 9 9 5.461 5.606 6 . 0 8 5 0.348 0.363 0 . 3 2 5 1.660 1 . 2 3 5 1.436
LSA 0 2.018 2.054 2 . 0 9 9 −0.291 0.149 0 . 5 0 5 0.371 0.396 0 . 3 2 8 1.416 1 . 0 7 3 1.096
  5 2.388 2.497 2 . 5 1 3 2.334 2.992 3 . 3 4 4 0.326 0.368 0 . 3 1 8 1.857 1 . 1 8 1 1.242
  10 2.715 2.861 2 . 8 7 0 5.378 5.948 6 . 3 4 6 0 . 2 7 1 0.329 0.316 2.331 1 . 4 0 6 1.528
  1. Bold values denote the best performance