Observing that most variations in pronunciation are strongly speaker and speaking style dependent, and that the introduction of pronunciation variants in a speaker-independent recognition system is of limited success, we refrain from applying multiple pronunciation variants in the speakerindependent case and instead introduce pronunciation variants without supervision when specializing the reco...

For a tonal language such as Mandarin Chinese, accurate pronunciation of tones is critical to meaning, and research suggests that computer-based programs that allow for visualization of pitch contours are helpful for improving learners’ pronunciation (Chun, 1998; Hardison, 2004; Levis & Pickering, 2004). This paper reports on a pilot study using speech analysis software (Praat), which allowed L...

Daniel J. Olson, Purdue University While a growing body of research has established the benefits of pronunciation training on second language (L2) production, these benefits have yet to be incorporated into the general skills language classroom in a systematic manner. Furthermore, although relatively new speech analysis software has been shown to be useful in providing visual feedback for L2 su...

Pronunciation dictionaries are the interface between orthographic and phonetic representation of the speech signal and are thereby a substantial component of speech recognition systems. In many systems simple canonical pronunciation forms are used within the dictionary. They represent the “correct” pronunciation as they are found in lexicons and neither contain the most frequent pronunciation n...

Modelling of pronunciation variability is an important part of the acoustic model of a speech recognition system. Good pronunciation models contribute to the robustness and portability of a speech recogniser. Usually pronunciation modelling is associated with the recognition lexicon which allows a direct control of HMM selection. However, in state-of-the-art systems the use of clustering techni...

Calculating posterior probability within a standard pronunciation space (SPS) is a common method in automatic pronunciation error detection (APED). However, to pronunciation errors outside the SPS, this kind of methods can only give an approximate solution, that may be not right in many applications. This paper expands the SPS to include more pronunciation errors, introduces a Bhattacharyya dis...

The primary data on pronunciation variation – e.g., dialect atlas data – is often recorded incommensurably, i.e. in different ways in different atlases, and even in different ways within the same atlas when teams of fieldworkers and transcribers are involved. In particular these data collections differ in the detail in which pronunciations are recorded, using between 40 and 100 different basic ...

We describe an algorithm to learn word pronunciations from acoustic data. The algorithm jointly optimizes the pronunciation of a word using (a) the acoustic match of this pronunciation to the observed data, and (b) how “linguistically reasonable” the pronunciation is. Variations of word pronunciations in the recognition dictionary (which was created by linguists), are used to train a model of w...

The present study investigated whether phonological information is activated automatically and, if so, how it affects handwritten production of Chinese characters. The form preparation paradigm was adopted. In the homogeneous blocks, target characters shared the first orthographic component and the pronunciation (Experiment 1), shared the first orthographic component only (Experiment 2), or sha...