PFHTS-IDSS: A Hybrid HTS-based Framework for Indonesian Speech Synthesis via Phoneme and Full-context Lab

In recent years, globalization has highlighted the importance of having machines that can truly provide customized communication for different languages. Majority of the research in the field focus on developing technologies for widely used languages such as English. In this study, we apply HMM-based speech synthesis (HTS) technology for Indonesian language. The proposed hybrid HTS-based framework, PFHTS-IDSS, uses phoneme and full-context lab to synthesize Indonesian with higher accuracy. First, we identify a list of Indonesian phonemes according to the initial-final structure of Chinese language. Based on this, we add zero-initials that match the Indonesian acoustic performance and HTS, which can make the synthesized speech natural and smooth. Second, we consider Indonesian phonemes as synthetic units to synthesize speech through the triphone and full-context lab. In addition, we design context properties of the full-context lab and the corresponding question set to train the acoustic model, which can eliminate machine sounds. Experimental results suggest that the accuracy of phoneme segmentation (PSA) and the naturalness of speech synthesis (SSN) are significantly improved via PFHTS-IDSS. Especially, the PSA of selecting phonemes as synthetic units reaches 88.3% and the corresponding SSN based on full-context lab is 4.1. The results demonstrated by PFHTS-IDSS presented in this paper may be used in multilingual free interactive system to promote better communication in terms of voice navigation, intelligent speaker and question-answering system.

Effects of word boxes on phoneme segmentation, word identification, and spelling for a sample of children with autism

The purpose of this study was to examine the effects of word boxes on the phoneme segmentation, word identification, and spelling performance of a sample of children with autism. Three children with autism were selected on the basis of similar performance on early literacy skills as measured by the Dynamic Indicators of Basic Early Literacy Skills (DIBELS) screening instrument. The word boxes is a method that involves students placing plastic letters into respective divided sections of a drawn rectangle (i.e., boxes) as each sound in a word is articulated. This method is designed to help children acquire phonological decoding skills. A multiple baseline design across literacy skills was employed to study the effects of word boxes on phoneme segmentation, word identification, and spelling. This study is important, as it was the first to examine the effects of this method with students with autism. Results suggested that all students showed increases in phoneme segmentation and word identification, with two of the students showing some improvement in spelling. Limitations and implications for future research and practice are discussed.

Predictive Validity of Kindergarten Progress Monitoring Measures Across the School Year: Application of Dominance Analysis

Although several measures are available for monitoring kindergarten reading progress, little research has directly compared them to determine which are superior in predicting year-end reading skills relative to other measures, and how validity may change across the school year as reading skills develop. A sample of 426 kindergarten students who were considered to be at risk for reading difficulty at the start of kindergarten were monitored across the year with a set of paper-based progress monitoring measures and a computer-adaptive test. Dominance analyses were used to determine the extent to which each measure uniquely predicted year-end reading skills relative to other measures. Although the computer-adaptive test was the most dominant predictor at the start of the year over letter sound fluency, letter naming fluency, and phoneme segmentation fluency, letter sound fluency was most dominant by December. Measures of fluency reading real words administered across the second half of the year were dominant to all other assessments. The implications for measure selection are discussed.

How African American English-Speaking First Graders Segment and Rhyme Words and Nonwords With Final Consonant Clusters

Purpose This study explored how typically developing 1st grade African American English (AAE) speakers differ from mainstream American English (MAE) speakers in the completion of 2 common phonological awareness tasks (rhyming and phoneme segmentation) when the stimulus items were consonant–vowel–consonant–consonant (CVCC) words and nonwords. Method Forty-nine 1st graders met criteria for 2 dialect groups: AAE and MAE. Three conditions were tested in each rhyme and segmentation task: Real Words No Model, Real Words With a Model, and Nonwords With a Model. Results The AAE group had significantly more responses that rhymed CVCC words with consonant–vowel–consonant words and segmented CVCC words as consonant–vowel–consonant than the MAE group across all experimental conditions. In the rhyming task, the presence of a model in the real word condition elicited more reduced final cluster responses for both groups. In the segmentation task, the MAE group was at ceiling, so only the AAE group changed across the different stimulus presentations and reduced the final cluster less often when given a model. Conclusion Rhyming and phoneme segmentation performance can be influenced by a child's dialect when CVCC words are used.

Phonological awareness errors mirror underlying phonological representations: Evidence from Hebrew L1 – English L2 adults

This article examines the effect of phonological context (singleton vs. clustered consonants) on full phoneme segmentation in Hebrew first language (L1) and in English second language (L2) among typically reading adults (TR) and adults with reading disability (RD) ( n = 30 per group), using quantitative analysis and a fine-grained analysis of errors. In line with earlier findings, overall mean scores revealed significant differences between the two groups. However, no qualitative differences were found. In both groups and languages, full phoneme segmentation overall scores for CVC stimuli were higher than CCVC stimuli. This finding does not align with previous findings, obtained from a phoneme isolation task, showing that isolation from a cohesive CV unit is the most difficult. A fine-grained analysis of errors was conducted to glean insight into this finding. The analysis revealed a preference for creating and preserving CV units in phoneme segmentation in both L1 and L2. This is argued to support the cohesion of the CV unit. The article argues that the effect of language-specific sub-syllabic representations on phonemic analysis may not be always observed in overall scores, yet it is reflected in specific patterns of phonological segmentation errors.

Phoneme Segmentation of Tamil Speech Signals Using Spectral Transition Measure

Process of identifying the end points of the acoustic units of the speech signal is called speech segmentation. Speech recognition systems can be designed using sub-word unit like phoneme. A Phoneme is the smallest unit of the language. It is context dependent and tedious to find the boundary. Automated phoneme segmentation is carried in researches using Short term Energy, Convex hull, Formant, Spectral Transition Measure(STM), Group Delay Functions, Bayesian Information Criterion, etc. In this research work, STM is used to find the phoneme boundary of Tamil speech utterances. Tamil spoken word dataset was prepared with 30 words uttered by 4 native speakers with a high quality microphone. The performance of the segmentation is analysed and results are presented.