D / Naming pictures of polysyllabic words (three or more syllables (PSWs)) seems to provide speech pathologists with information about communication status not necessarily present when naming pictures of short words (monosyllabic words (MSWs) and di-syllabic words (DSWs)). Typically developing children and children with speech, language and literacy impairments err on PSWs even when short words are accurate. In this study, typical behaviour of PSW production was delimited and a model of PSW acquisition was developed because if erroneous PSWs mark impairment, then circumscribing the tolerances of them in typically developing speech is necessary to differentiate it from impairment. A proportional stratified, cluster sampling procedure was used to locate 354 children, aged 3;0 to 7;11 years, of whom 283 met the selection criteria, including normal hearing, language and cognition. All English phonemes were repeatedly sampled in 166 words, elicited through picture naming, that were varied for syllable number, stress and shape. Syllable, age and interaction effects were present with more mismatches in PSWs than in short words, decreasing with increasing age. Mismatches were captured in five a priori patterns of deletions, additions and reordering of syllables and segments in words as well as alterations of consonants or vowels in words that preserved the phonotactic shape. However, as all five patterns were word-specific, each affecting a core group of words containing PSWs and DSWs, the syllable effect was modified. It appeared to be a proxy for a complex interaction between segmental and prosodic features common to the core words that included non-final weak syllables, within-word consonant sequences that required labial-velar movements, velar and sonorant sounds and sounds that shared place or manner features, severally or together. The production changes conformed to the predictions of the model of PSW acquisition. These changes reflected alterations in the phonological representation, motor planning and motor execution skills aspects of the speech processing system. The phonological representation, changing from holistic to fine-grained, was argued as the key change because information for motor planning and execution was liberated that culminated in increased accuracy. If children’s productions of the PSWs used in this study exceed the tolerances defined in this thesis, impairment may be indicated. Future research is needed to determine that possibility.

D / Naming pictures of polysyllabic words (three or more syllables (PSWs)) seems to provide speech pathologists with information about communication status not necessarily present when naming pictures of short words (monosyllabic words (MSWs) and di-syllabic words (DSWs)). Typically developing children and children with speech, language and literacy impairments err on PSWs even when short words are accurate. In this study, typical behaviour of PSW production was delimited and a model of PSW acquisition was developed because if erroneous PSWs mark impairment, then circumscribing the tolerances of them in typically developing speech is necessary to differentiate it from impairment. A proportional stratified, cluster sampling procedure was used to locate 354 children, aged 3;0 to 7;11 years, of whom 283 met the selection criteria, including normal hearing, language and cognition. All English phonemes were repeatedly sampled in 166 words, elicited through picture naming, that were varied for syllable number, stress and shape. Syllable, age and interaction effects were present with more mismatches in PSWs than in short words, decreasing with increasing age. Mismatches were captured in five a priori patterns of deletions, additions and reordering of syllables and segments in words as well as alterations of consonants or vowels in words that preserved the phonotactic shape. However, as all five patterns were word-specific, each affecting a core group of words containing PSWs and DSWs, the syllable effect was modified. It appeared to be a proxy for a complex interaction between segmental and prosodic features common to the core words that included non-final weak syllables, within-word consonant sequences that required labial-velar movements, velar and sonorant sounds and sounds that shared place or manner features, severally or together. The production changes conformed to the predictions of the model of PSW acquisition. These changes reflected alterations in the phonological representation, motor planning and motor execution skills aspects of the speech processing system. The phonological representation, changing from holistic to fine-grained, was argued as the key change because information for motor planning and execution was liberated that culminated in increased accuracy. If children’s productions of the PSWs used in this study exceed the tolerances defined in this thesis, impairment may be indicated. Future research is needed to determine that possibility.

Undoubtedly, state-of-the-art unit selection-based concatenative speech systems produce very high quality synthetic speech. this is due to a large speech database containing many instances of each speech unit, with a varied and natural distribution of prosodic and spectral characteristics. the join cost, which measures how well two units can be joined together is one of the main criteria for selecting appropriate units from this large speech database. The ideal join cost is one that measures percieved discontinuity based on easily measurable spectral properties of the units being joined, inorder to ensure smooth and natural sounding synthetic speech. During first part of my research, I have investigated various spectrally based distance measures for use in computation of the join cost by designing a perceptual listening experiment. A variation to the usual perceptual test paradigm is proposed in this thesis by deliberately including a wide range of qualities of join in polysyllabic words. The test stimuli are obtained using a state-of-the-art unit-selection text-to-speech system: rVoice from Rhetorical Systems Ltd. Three spectral features Mel-frequency cepstral coefficients (MFCC), line spectral frequencies (LSF) and multiple centroid analysis (MCA) parameters and various statistical distances - Euclidean, Kullback-Leibler, Mahalanobis - are used to obtain distance measures. Based on the correlations between perceptual scores and these spectral distances. I proposed new spectral distance measures, which have good correlation with human perception to concatenation discontinuities. The second part of my research concentrates on combining join cost computation and the smoothing operation, which is required to disguise joins, by learning an underlying representation from the acoustic signal. In order to accomplish this task, I have chosen linear dynamic models (LDM), sometimes known as Kalman filters. Three different initialisation schemes are used prior to Expectation-Maximisation (KM) in LDM training. Once the models are trained, the join cost is computed based on the error between model predictions and actual observations. Analytical measures are derived based on the shape of this error plot. These measures and initialisation schemes are compared by computing correlations using the perceptual data. The LDMs are also able to smooth the observations which are then used to synthesise speech. To evaluate the LDM smoothing operation, another listening test is performed where it is compared with the standard methods (simple linear interpolation). I have compared the best three join cost functions, chosen from the first and second parts of my research, subjectively using a listening test in the third part of my research. in this test, I also evaluated different smoothing methods: no smoothing, linear smoothing and smoothing achieved using LDMs.