Grapheme-to-phoneme conversion system for English is being developed for further integration into speech synthesis system within TC-STAR project. In this work we describe experiments performed using two different machine learning techniques. The pronunciation was predicted both for stressed and unstressed lexicon and the results were compared. Analysis of different parameters that may influence...

We introduce a new language-independent procedure for grapheme-phoneme conversion, syllabification, and word stress assignment. Grapheme-phoneme conversion and syllabification is carried out by means of fallback sequences of decision trees trained on varying context sizes. Word stress is determined within an analogy-based framework by means of a Bayes classifier. Evaluation results on six langu...

Prior research has purported to show that words with infrequent phoneme-grapheme correspondences are more difficult to spell than words with frequent phoneme-grapheme correspondences. Defining exactly what a phoneme-grapheme relationship is, however, is not necessarily straightforward. There are a number of different assumptions that can be made. In this study, we developed four metrics of soun...

This paper compares the recognition accuracy of a phonemebased automatic speech recognition system with that of a grapheme-based system, using Afrikaans as case study. The first system is developed using a conventional pronunciation dictionary, while the latter system uses the letters of each word directly as the acoustic units to be modelled. We ensure that the pronunciation dictionary we use ...

This study shows that biases exist in the associations of letters with colours across individuals both with and without grapheme-colour synaesthesia. A group of grapheme-colour synaesthetes were significantly more consistent over time in their choice of colours than a group of controls. Despite this difference, there were remarkable inter-subject agreements, both within and across participant g...

In double naming tasks requiring the production of consecutive spoken and written responses to the same picture, subject ECA produced inconsistent lexical responses in the say-then-write (stimulus: organ; spoken response: “church;” written response: piano) but not in the write-then-say condition (organ ® piano ® “piano”). This observation, together with the fact that ECA had damage to the seman...

Synaesthesia is a condition in which one property of a stimulus triggers a secondary experience not typically associated with the first (e.g., seeing achromatic graphemes can evoke the perception of color). Recent work has explored a variety of cognitive and perceptual traits associated with synaesthesia. One example is in the domain of personality, where higher rates of positive schizotypy and...

We aim to develop a Speech-to-Speech translation system between Modern Standard Arabic and Algiers dialect. Such a system must include a Text-to-Speech module which itself must include a Grapheme-to-Phoneme converter. Algiers dialect is an Arabic dialect concerned by the most problems of Modern Standard Arabic in NLP area. Furthermore, it could be considered as an under-resourced language becau...