Search

This chapter deals with the estimation formant estimation and tracking of the movements of the spectral resonances of human formant vocal tracts, also known as formants. The representation or modeling of speech in terms of formants is useful in several areas of speech processing: coding, recognition, synthesis, and enhancement, as formants efficiently describe essential aspects of speech using a very limited set of parameters. However, estimating formants is more difficult than simply searching for peaks in an amplitude spectrum, as the spectral peaks of vocal-tract output depend upon a variety for factors in complicated ways: vocal-tract shape, excitation, and periodicity. We describe in detail the formal task of formant tracking, and explore its successes and difficulties, as well as giving reasons for the various approaches. [ABSTRACT FROM AUTHOR]

This paper presents a hybrid technique combining the Karhonen-Loeve Transform (KLT), the Multilayer Perceptron (MLP) and Genetic Algorithms (GAs) to obtain less-variant Mel-frequency parameters. The advantages of such an approach are that the robustness can be reached without modifying the recognition system, and that neither assumption nor estimation of the noise are required. To evaluate the effectiveness of the proposed approach, an extensive set of continuous speech recognition experiments are carried out by using the NTIMIT telephone speech database. The results show that the proposed approach outperforms the baseline and conventional systems. [ABSTRACT FROM AUTHOR]

The article examines ways that computers use to analyze speech and to simulate production or perception of speech signals. There are many applications for speech processing: text-to-speech synthesis, speech recognition, speaker verification, and speech coding. All these uses require converting the original speech into a compact digital form, or vice versa, and related digital signal processing. The objective of speech analysis is to transform speech samples into a more useful information representation, for specific objectives other than simply furnishing speech to human ears.

The article focuses on speech processing. Speech processing refers to the manipulation or transformation of speech signals, mostly by computer, to more useful forms of information. The first step in speech analysis is to convert the continuously- varying speech waves into a bit stream, so that computers may receive and process the speech signal. For speech recognition, much is known about the speech production process linking a text and its phonemes to the spectra and intonation of a corresponding speech signal. Diphone synthesizers try to circumvent this problem by storing the parameter transitions from one phoneme to the next since co-articulation primarily influences only the immediately adjacent phonemes.