Skip to main content
. 2023 May 11;18(5):e0285629. doi: 10.1371/journal.pone.0285629

Table 9. WERs for different SE algorithms.

Noisy Speech DWAtten-LSTM LSTM LSTM-KF LMMSE
46.33% 13.11% 15.73% 14.97% 27.68%