![]() The main contributions are comprehensive evaluations of the effect of both additive white Gaussian noise (AWGN) and non-stationary noise (NSN) (with and without a G.712 type handset) upon identification performance. Normalization is applied by employing cepstral mean and variance normalization (CMVN) and feature warping (FW), together with acoustic modeling using a Gaussian mixture model-universal background model (GMM-UBM). In this study, a speaker identification system is considered consisting of a feature extraction stage which utilizes both power normalized cepstral coefficients (PNCCs) and Mel frequency cepstral coefficients (MFCC). The Praat software was also used to verify the measures of jitter in the synthesized voice signals. It is showed that jitter could be obtained using the model proposed. Some samples of synthesized voices in these cases are obtained. The probability density function of the fundamental frequency related to the voice signals produced are constructed and compared for different levels of jitter. ![]() The corresponding stiffness of each vocal fold is considered as a stochastic process, and its modeling is proposed. ![]() Large values for jitter variations can indicate a pathological characteristic of the voice. The jitter has been the subject for researchers due to its important applications such as the identification of pathological voices (nodules in the vocal folds, paralysis of the vocal folds, or even, the vocal aging, among others). The objective of this paper is the construction of a stochastic model for jitter using a one-mass mechanical model of the vocal folds, which assumes complete right-left symmetry of the vocal folds, and which considers motions of the vocal folds only in the horizontal direction. The observation of the glottal cycles variations suggests that jitter is a random phenomenon described by random deviations of the glottal cycle lengths in relation to a corresponding mean value and, in general, its values are expressed as a percentage of the duration of the glottal pulse. The quasiperiodic oscillation of the vocal folds causes perturbations in the length of the glottal cycles, which are known as jitter. Glottal signals and voice signals are generated with jitter and the probability density function of the fundamental frequency is constructed for several values of the hyperparameters that control the level of jitter. The stiffnesses taken into account in the model are considered as stochastic processes and their modeling are proposed. The aim of this paper is to construct a stochastic model of jitter using a two-mass mechanical model of the vocal folds, assuming complete right-left symmetry of the vocal folds and considering the motion of the vocal folds only in the horizontal direction. Its study has been developed due to important applications such as aid in identification of voices with pathological characteristics, when its values are large, because a normal voice has naturally a low level of jitter. It can be modeled as a random phenomenon described by the deviations of the glottal cycle length in relation to a mean value. Jitter is a phenomenon caused by the perturbation in the length of the glottal cycles due to the quasi-periodic oscillation of the vocal folds in the production of the voice.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |