The Gaussian mixture model (GMM) method is popular and efficient for voice conversion (VC), but it is often subject to overfitting. In this paper, the principal component regression (PCR) method is adopted for the spectral mapping between source speech and target speech, and the numbers of principal components are adjusted properly to prevent the overfitting. Then, in order to better model the nonlinear relationships between the source speech and target speech, the kernel principal component regression (KPCR) method is also proposed. Moreover, a KPCR combined with GMM method is further proposed to improve the accuracy of conversion. In addition, the discontinuity and oversmoothing problems of the traditional GMM method are also addressed. On the one hand, in order to solve the discontinuity problem, the adaptive median filter is adopted to smooth the posterior probabilities. On the other hand, the two mixture components with higher posterior probabilities for each frame are chosen for VC to reduce the oversmoothing problem. Finally, the objective and subjective experiments are carried out, and the results demonstrate that the proposed approach shows greatly better performance than the GMM method. In the objective tests, the proposed method shows lower cepstral distances and higher identification rates than the GMM method. While in the subjective tests, the proposed method obtains higher scores of preference and perceptual quality.
Pyrolysis is potentially an effective treatment of oily sludge for oil recovery, and the addition of a catalyst is expected to affect its pyrolysis behavior. In the present study, Fe/Al-pillared bentonite with various Fe/Al ratios as pyrolysis catalyst is prepared and characterized by XRD, N2 adsorption, and NH3-TPD. The integration of Al and Fe in the bentonite interlayers to form pillared clay is evidenced by increase in the basal spacing. As a result, a critical ratio of Fe/Al exists in the Fe/Al-pillared bentonite catalytic pyrolysis for oil recovery from the sludge. The oil yield increases with respect to increase in Fe/Al ratio of catalysts, then decreases with further increasing of Fe/Al ratio. The optimum oil yield using 2.0 wt% of Fe/Al 0.5-pillared bentonite as catalyst attains to 52.46% compared to 29.23% without catalyst addition in the present study. In addition, the addition of Fe/Al-pillared bentonite catalyst also improves the quality of pyrolysis-produced oil and promotes the formation of CH4. Fe/Al-pillared bentonite provides acid center in the inner surface, which is beneficial to the cracking reaction of oil molecules in pyrolysis process. The present work implies that Fe/Al-pillared bentonite as addictive holds great potential in industrial pyrolysis of oily sludge.
A novel VC (voice conversion) method based on hybrid SVR (support vector regression) and GMM (Gaussian mixture model) is presented in the paper, the mapping abilities of SVR and GMM are exploited to map the spectral features of the source speaker to those of target ones. A new strategy of F0 transformation is also presented, the F0s are modeled with spectral features in a joint GMM and predicted from the converted spectral features using the SVR method. Subjective and objective tests are carried out to evaluate the VC performance; experimental results show that the converted speech using the proposed method can obtain a better quality than that using the state-of-the-art GMM method. Meanwhile, a VC method based on non-parallel data is also proposed, the speaker-specific information is investigated using the SVR method and preliminary subjective experiments demonstrate that the proposed method is feasible when a parallel corpus is not available.