Estimation of the signal-to-noise ratio with amplitude modulation spectrograms

An algorithm is proposed which automatically estimates the local signalto-noise ratio (SNR) between speech and noise. The feature extraction stage of the algorithm is motivated by neurophysiological findings on amplitude modulation processing in higher stages of the auditory system in mammals. It analyzes information on both center frequencies and amplitude modulations of the input signal. This information is represented in two-dimensional patterns, so-called Amplitude Modulation Spectrograms (AMS). A neural network is trained on a large number of AMS patterns generated from mixtures of speech and noise. After 1A modified version of this Chapter has been submitted to Speech Communication: Tchorz and Kollmeier (2000) "Estimation of the signal-to-noise ratio with amplitude modulation spectrograms". 7 

• Type: Journal
• Year: 2002
• Where: SPEECH
• Authors: Jürgen Tchorz, Birger Kollmeier
• Keywords: Amplitude Modulation, Amplitude Modulation Spectrograms, So-called Amplitude Modulation, SPEECH 2002