How does the automatic phonemic transcription work?

The draft phonemic transcription is basically obtained by linking the phonemic transcription of the words in isolation. However, you have to take the following into account:
  • Words not found in the database are left as they are, but between '<' and '>'.
    getafe > <getafe>
  • The following symbols are omited from the original text: ':', '?', '!', '(' and ')'
    how are you? > haʊ ə ju
  • The tool is not case-sensitive. It doesn't distinguish lower case from upper case. Therefore, "abc" is not considered different from "ABC".
    abc > ˌeɪbi:ˈsi:
    ABC > ˌeɪbi:ˈsi:
  • Homographs. Some homographs, words with same spelling, have different pronunciations. PhoTransEdit shows different pronunciations if the word is recognised as an homograph in the database. To distinguish the right pronunciation the words around the homograph are used.
    a story > ə ˈstɔ:ri
    letter a > ˈletər
    an articulate speaker > ən ɑ:ˈtɪkjʊlət ˈspi:kə
    to articulate an idea > tu: ɑ:ˈtɪkjuleɪt ən aɪˈdɪə
  • Strong and weak forms. The weak forms are generally used. The forms are always strong at the end of phrases, except if it is a pronoun.
    I found some > aɪv faʊnd sʌm
    I found some coins > aɪv faʊnd səm koɪnz
    There are exceptions:
    the > ðə before a consonant sound (the boy, the house) and ði before a vowel sound (the egg, the hour)
    to > before a consonant sound and tu before a vowel sound
    is > ɪz after a word ending in s, z, ʃ, ʒ, ʧ, ʤ. The contracted form s is used after a word ending in p, t, k, f, θ while z is used after one ending in a vowel sound or b, d, g, v, ð, m, n, ŋ, l and AmE r
    Do not include a carriage return or a full stop in the middle of a sentence. They are considered the beginning of a new sentence. Note: Weak and strong forms are determined by stress, intonation and phonetic context, therefore it is difficult to establish an automated rule that works for all cases.
  • Stress in compounds. The pronunciation of compound words may differ from the concatenation of their component parts. PhoTransEdit shows the pronunciation of compound words included in its database.
    acid > ˈæsɪd
    acid rain > ˌæsɪd ˈreɪn
  • Some dictionaries show the symbol (r) in the pronunciation of some words. For example, far /fɑ:(r)/. This means that far will be pronounced /fɑ:r/ when followed by a vowel sound and /fɑ:/ in all other cases.
    far country > fɑ: ˈkʌntri
    far away > fɑ:r əˈweɪ
  • Numbers. Numbers are converted into words and then transcribed.
    101 > wʌn ˈhʌndrəd ənd wʌn
    XXIV > ˈtwentɪ fɔːr
    These rules are only followed by the words included in the database. Let me know of any homograph or any other word not recognised.