Automatic discrimination of speech and music is an important tool in many multimedia applications. Previous work has focused on using long-term features such as differential parameters, variances, and time-averages of spectral parameters. These classifiers use features estimated over windows of 0.5-5 seconds, and are relatively complex. In this paper, we present our results of combining the line spectral frequencies (LSFs) and zero-crossing-based features for frame-level narrowband speech/music discrimination. Our classification results for different types of music and speech show the good discriminating power of these features. Our classification algorithms operate using only a frame delay of 20 ms, making them suitable for real-time multimedia applications.
[1] The linear and nonlinear morphological behavior of double-barred coastal systems under the forcing of obliquely incident waves is studied using a nonlinear numerical model. The linearly most unstable bed forms consist of crescentic patterns (rip channels), whose spacing depends on the magnitude of the longshore current velocity. Using the nonlinear model, six morphodynamic experiments have been performed with various initial bed perturbations in order to assess, among others, the influence of the initial bed perturbation on the morphodynamic evolution. The nonlinear experiments have been pursued well into the nonlinear regime, showing that after a phase of initial exponential growth, a highly dynamic behavior is observed and no equilibrium state is reached. The spacings predicted with the linear stability analysis are observed during the exponential phase of the nonlinear experiments. In the dynamic phase, however, four to seven modes significantly contribute to the resulting bed features. In this final stage, the apparent wavelength of 1000 m of the resulting bed forms on the inner bar is quite insensitive to the initial bed perturbation. On the outer bar it seems that the longer the wavelength of the initial bed perturbation, the longer the wavelength of the final bed forms in the dynamic phase and the larger the migration celerity. In general, the bed forms can be characterized as crescentic or undulating bed patterns. Good correspondence between simulated and observed spacings, shapes and migration celerities are found.
[1] The initial growth of bed perturbations on planar sloping beaches under the forcing of obliquely incident, breaking waves is investigated using a state-of-the-art numerical model. This allows for a systematic investigation of the sensitivity of the spatial structures of the bed perturbations and their growth and migration rates to different model formulations and parameterizations. If the sediment is only transported in the direction of the net current velocity and sediment stirring is taken proportional to the wave height squared, growing up-current oriented crescentic bars are found with a preferred spacing of 800 m and a down-current migration rate of 10 m d À1 . Varying the angle of wave incidence, drag coefficient and bed slope results in qualitatively similar growing bed forms. Using an Engelund and Hansen transport formula, very oblique down-current oriented bars are obtained that grow in time. No preferred wavelength, however, is found. Using the Bailard transport formula results in growing, up-current oriented bars with a preferred spacing smaller than 300 m for wave angles smaller than 7°. When using either the Engelund and Hansen or Bailard sediment transport formulation, it is essential to take the transport in the direction of the wave orbital velocity into account in order to have growing bed perturbations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.