Abstract not found

Abstract not found

Support vector machines (SVMs) can be trained to detect manner transitions between phones and to identify the manner and place of articulation of any given phone. The SVMs can perform these tasks with high accuracy using a variety of acoustic representations. The SVMs generalize well to unseen test data if these data were created under identical conditions to the training corpus. Unseen acoustic data from different corpora present a problem for the SVM, even if these acoustic data were generated under similar conditions. The discriminant outputs of these SVMs are used to create both a hybrid SVM/HMM (hidden Markov model) phone recogni-tion system and a hybrid SVM/HMM word recognition system. There is a significant improvement in both phone and word recognition accuracy when these SVM discrim-inant features are used instead of mel frequency cepstral coefficients (MFCCs).

Keywords: 1. Ii8 ODUCTIl It is twelfth-century Japan, and a nobleman has been killed.

In this paper work, past, present and future, in articulatory synthesis applied to Portuguese is presented .

Phonetic variation as found in various speech styles is a rich area for research on spoken word recognition. Research on spoken word recognition has focused on careful, easily controlled speech styles. This dissertation investigates the processing of the American English Flap. Specifically, it focuses on the effect of reduction on processing. The main question asked in this dissertation is whether listeners adjust their expectations for how segments are realized based on speech style. Even more broadly, how do listeners process or recognize reduced speech? Two specific questions are asked that address individual parts of the broad question. First, how does reduction affect listeners’ recognition of words? Is it more difficult for listeners to recognize words pronounced in reduced forms, or is it perhaps easier for listeners to recognize reduced forms? Second, do listeners adjust their expectations about reduction based on preceding speech style (context)? Four experiments were designed using the auditory lexical decision and cross-modal identity priming tasks. Listeners’ responses to reduced and unreduced flaps (e.g. unreduced [pʌɾl̩] as opposed to reduced [pʌɾ̞l̩]) were recorded. The results of this work show that the phonetic variation found in speech styles containing reduction causes differences in processing. Processing of reduced speech is inhibited by weakened acoustic information or mismatch to the underlying phonemic representation in the American English flap. Listeners use information about speech style to process the widely varying acoustic reflections of a segment in connected speech. The implications of these findings for models of spoken word recognition are discussed.

Support vector machines (SVMs) are trained to detect acoustic-phonetic landmarks, and to identify both the manner and place of articulation of the phones producing each landmark with high accuracy. The discriminant outputs of these SVMs are used as input features for a standard HMM based ASR system. There is a significant improvement in both the phone and word recognition accuracy when using these SVM discriminant features when compared to the phone and word recognition accuracy of an MFCC based recognizer.