Overview of the proposed model

  

Figure: Overview of signal-flow in the proposed model.

An overview of signal processing of the proposed model is shown in Fig. . First, noisy vowel /a/ f(t) shown in Fig. A (the SNR of f(t)is 10 dB) is decomposed into S_k(t) and $\phi_k(t)$ as shown in Figs. B and C, respectively. Next, F₀(t) is estimated as shown in Fig. D. The concurrent time-frequency region of the desired signal f₁(t) is determined using constraints (i) and (iii) as shown in Figs. E and F. Finally, the instantaneous amplitudes and the instantaneous phases of the two signals are determined from S_k(t) and $\phi_k(t)$ using constraints (ii) and (iv). The determined A_k(t) and $\theta_{1k}(t)$ are shown in Figs. H and I, respectively. The segregated signal $\hat{f}_1(t)$ is shown in Fig. J. In this figure, note that the segregated B_k(t), $\theta_{2k}(t)$, and $\hat{f}_2(t)$ are omitted.