Skip to main content
. 2017 Jun 23;141(6):4705–4714. doi: 10.1121/1.4986931

FIG. 5.

FIG. 5.

Comparison of the DNN and LSTM in terms of STOI improvement (in %) with the unseen babble noise. (a) Performance of the DNN and LSTM on six seen speakers at −5 dB SNR. (b) Performance of the DNN and LSTM on six unseen speakers at −5 dB SNR. (c) Performance of the DNN and LSTM on six seen speakers at −2 dB SNR. (d) Performance of the DNN and LSTM on six unseen speakers at −2 dB SNR.