Research Article
Deep Learning-Based Amplitude Fusion for Speech Dereverberation
Table 4
Average scores of various speech enhancement methods.
| | Type | Method | PESQ | SNR |
| | — | MIX | 1.48 | 3.62 | | — | WPE | 1.60 | 4.23 | | Mapping | MAP | 1.77 | 7.59 |
| | Masking | cIRM | 1.63 | 7.18 | | PSM | 1.80 | 6.93 | | TDR | 1.77 | 7.76 | | I_IRM | 1.54 | 4.02 | | IAM_A | 1.85 | 7.71 | | DCC_A | 1.83 | 7.64 | | IRM | 1.57 | 4.24 | | IAM | 1.91 | 7.17 | | DCC | 1.86 | 8.21 |
| | Masking + masking | IAM + IRM | 1.74 | 5.91 | | IRM + DCC | 1.79 | 6.66 | | IAM + DCC | 1.90 | 7.89 |
| | Mapping + masking | MAP + I_IRM | 1.90 | 7.30 | | MAP + IRM | 1.91 | 7.14 | | MAP + IAM | 2.00 | 8.19 | | MAP + DCC | 2.01 | 8.78 |
|
|
