Fig. 3From: Environment-dependent denoising autoencoder for distant-talking speech recognitionTwo-step environment-dependent DAEBack to article page