Fig. 1From: A priori SNR estimation and noise estimation for speech enhancementsp04.wav added by 0-dB babble noise. a Spectrogram of clean speech. b Spectrogram of noisy speech c when normalized frequency point k = 20, d when normalized frequency point k = 60, and e when normalized frequency point k = 120Back to article page