Model B: the power spectrum model of masking

  

Figure: Model B: a power spectrum model of masking.

In the power spectrum model [1], it is assumed that when a listener is trying to detect a sinusoidal signal with a particular center frequency in a background noise, he uses of the output of a single auditory filter whose center frequency is close to the signal frequency, and which has the highest signal-to-masker ratio. Therefore, it can be considered that the component passed through the single auditory filter only affects masking. In particular the masking threshold for a sinusoidal signal is determined by the amount of noise passing through the auditory filter.

The power spectrum model consists of Model B as shown in Fig. . The output of the auditory filter X_k(t) is one of the outputs of the auditory filterbank. This filter consists of the gammatone filter whose center frequency is 1 kHz and bandwidth is 1 ERB. In this model, the sinusoidal signal $\hat{f}_{1,{\rm {B}}}(t)$ extracted from the masked signal f(t) is the output of the single auditory filter X_k(t).