Formulating a phonological representation system

Finally, using a Unear phonemic representation has the benefit that it makes automatic data-base labelling considerably easier. In Chapter 17 we shall consider the issue of labelling a speech database with phoneme units, both by hand and by computer. As we shall see, it is easier to perform this labelling if we assume a linear pronunciation model, because this works well with automatic techniques such as hidden Markov models. 

Having in principle chosen a system that uses phonemes and syllable structure, we now turn to the issue of how to define this for a given speaker, accent or language. It is essential that this is done with care if we decide, say, to use a single phoneme /ch/ instead of /t/ + /sh/, we have to use this consistently, snch that we represent church as /ch n r ch/, not as /t sh u r ch/ or /ch u r t sh/ and so on. 

It does seem somewhat wasteful that separate pronunciation systems and therefore lexicons are required for British English and American English while differences of course exist, pronunciations in the two accents aren t completely different. Fitt [161], [162], [163] proposed a solution to this whereby a more-abstract pronunciation drawn from a large set of phonemes was used as a base lexicon. From this, filters could be used to generate accents for any accent of English. 

---