This paper describes the methods of emphasizing spectral cues from the sound waveforms polluted with colored noise. The ambiguous spectral features of noisy speech data are emphasized on cepstral domain. Several methods for this purpose are defined. Some methods are based on spectral subtraction method. Others are based on the emphasis of spectral inclination by the nature of cepstral coefficients. These methods are applied to the pattern recognition of noisy speech. The following describes the examination of the performance of these parameters and the usage of those parameters is discussed.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.