Figure 2From: A perceptual masking approach for noise robust speech recognition Tone and noise maskers for part of the utterance “three six seven” corrupted with subway noise of 5 dB SNR. Back to article page