English (India) MFA dictionary v3.0.0#
@techreport{mfa_english_india_mfa_dictionary_2024,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={English (India) MFA dictionary v3.0.0},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/English/English (India) MFA dictionary v3_0_0.html}},
year={2024},
month={Feb},
}
G2P models |
Installation#
Install from the MFA command line:
mfa model download dictionary english_india_mfa
Or download from the release page.
The dictionary available from the release page and command line installation has pronunciation and silence probabilities estimated as part acoustic model training (see Silence probability format and training pronunciation probabilities for more information. If you would like to use the version of this dictionary without probabilities, please see the [plain dictionary](https://raw.githubusercontent.com/MontrealCorpusTools/mfa-models/main/dictionary/english/mfa/English (India) MFA dictionary v3_0_0.dict).
Intended use#
This dictionary is intended for forced alignment of English transcripts.
This dictionary uses the MFA phone set for English, and was used in training the English MFA acoustic model. Pronunciations can be added on top of the dictionary, as long as no additional phones are introduced.
Performance Factors#
When trying to get better alignment accuracy, adding pronunciations is generally helpful, especially for different styles and dialects. The most impactful improvements will generally be seen when adding reduced variants that involve deleting segments/syllables common in spontaneous speech. Alignment must include all phones specified in the pronunciation of a word, and each phone has a minimum duration (by default 10ms). If a speaker pronounces a multisyllabic word with just a single syllable, it can be hard for MFA to fit all the segments in, so it will lead to alignment errors on adjacent words as well.
Ethical considerations#
Deploying any Speech-to-Text model into any production setting has ethical implications. You should consider these implications before use.
Demographic Bias#
You should assume every machine learning model has demographic bias unless proven otherwise. For pronunciation dictionaries, it is often the case that transcription accuracy and lexicon coverage for the prestige variety modeled in this dictionary compared to other variants. If you are using this dictionary in production, you should acknowledge this as a potential issue.
IPA Charts#
Consonants#
Obstruent symbols to the left of are unvoiced and those to the right are voiced.
Manner |
Labial |
Labiodental |
Dental |
Alveolar |
Alveopalatal |
Retroflex |
Palatal |
Velar |
Glottal |
---|---|---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 16,611 Examples: * murr: [m ɜ] * swamp: [s ʋ ɒ m p] * games: [ɡ eː m z] * small: [s m ɒː l] Occurrences: 3,628 Examples: * amira: [ə mʲ iː ɹ ə] * mulan: [mʲ ʊ l ə n] * humid: [ç ʉː mʲ ɪ ɖ] * musa: [mʲ ʉː z ə] |
Occurrences: 4 Examples: |
Occurrences: 34,120 Examples: * caine: [c eː n] * snag: [s n a ɡ] * icon: [aj k ə n] * jin: [dʒ ɪ n] |
Occurrences: 6,276 Examples: * runny: [ɹ ə ɲ i] * benin: [b ɛ ɲ iː n] * ensue: [ɛ ɲ s j ʉː] * natal: [ɲ eː ʈ ə l] |
Occurrences: 6,165 Examples: * junk: [dʒ ə ŋ k] * lanka: [l a ŋ k ə] * hank: [h a ŋ k] * swung: [s ʋ ə ŋ] |
||||
Stop |
Occurrences: 15,694 Examples: * pita: [p iː ʈ ə] * crap: [k ɹ a p] * picks: [p ɪ k s] * port: [p ɒː ʈ] Occurrences: 638 Examples: * tupi: [ʈ ʉː pʲ i] * puce: [pʲ ʉː s] * lippi: [ʎ ɪ pʲ i] * copy: [k ɒ pʲ i] Occurrences: 13 Examples: Occurrences: 10,938 Examples: * brian: [b ɾ aj ə n] * balch: [b a l k] * abegg: [ə b ɛ ɡ] * bon: [b ɒ n] Occurrences: 2,232 Examples: * beans: [bʲ iː n z] * bambi: [b a m bʲ i] * beaux: [bʲ oː] * beal: [bʲ iː l] |
Occurrences: 2,474 Examples: * thor: [t̪ ɒː] * thou: [t̪ aw] * path: [p ɑː t̪] * thee: [t̪ iː] Occurrences: 654 Examples: * thou: [d̪ aw] * meath: [mʲ iː d̪] * lithe: [l aj d̪] * width: [ʋ ɪ d̪] |
Occurrences: 29,086 Examples: * stoke: [s ʈ oː k] * front: [f ɾ ə n ʈ] * studs: [s ʈ ə ɖ z] * stack: [s ʈ a k] Occurrences: 7,548 Examples: * pity: [p ɪ ʈʲ i] * dante: [ɖ a n ʈʲ i] * steep: [s ʈʲ iː p] * roti: [ɹ oː ʈʲ i] Occurrences: 115 Examples: * tweed: [ʈʷ iː ɖ] * twigs: [ʈʷ ɪ ɡ z] * twain: [ʈʷ eː n] * twins: [ʈʷ ɪ n z] Occurrences: 23,240 Examples: * ahmad: [ɑː m ə ɖ] * dough: [ɖ oː] * gould: [ɡ ʊ ɖ] * zelda: [z ɛ l ɖ ə] |
Occurrences: 5,884 Examples: * mckim: [m ə c ɪ m] * kenny: [c ɛ ɲ i] * gorky: [ɡ ɒː c i] * wiki: [ʋ iː c i] Occurrences: 455 Examples: * equip: [ɪ cʷ ɪ p] * queen: [cʷ iː n] * query: [cʷ ɪ ɹ i] * quest: [cʷ ɛ s ʈ] Occurrences: 1,787 Examples: * get: [ɟ ɪ ʈ] * grain: [ɟ ɹ eː n] * agnes: [a ɟ n ɪ s] * rugby: [ɹ ə ɟ bʲ i] Occurrences: 83 Examples: * gwin: [ɟʷ ɪ n] * guido: [ɟʷ iː ɖ oː] * gwyn: [ɟʷ ɪ n] * gwynn: [ɟʷ ɪ n] |
Occurrences: 21,685 Examples: * ask: [ɑː k s] * axed: [a k s ʈ] * chas: [k a z] * risk: [ɾ ɪ s k] Occurrences: 481 Examples: * quake: [kʷ eː k] * quite: [kʷ aj] * quoll: [kʷ ɒ l] * quiet: [kʷ aj ɪ ʈ] Occurrences: 6,127 Examples: * guest: [ɡ ɛ s] * cogan: [k oː ɡ ə n] * glass: [ɡ l ɑː s] * exact: [ɪ ɡ z a ʈ] Occurrences: 67 Examples: * guo: [ɡʷ oː] * guam: [ɡʷ ɒ m] * guang: [ɡʷ ɑː ŋ] * guava: [ɡʷ ɑː ʋ ə] |
Occurrences: 26 Examples: * wyatt: [ʋ aj a ʔ] * fleet: [fʲ ʎ iː ʔ] * not: [n ɒ ʔ] * but: [b ə ʔ] |
|||
Affricate |
Occurrences: 3,275 Examples: * tract: [tʃ ɹ a k ʈ] * witch: [ʋ ɪ tʃ] * chop: [tʃ ɒ p] * chill: [tʃ ɪ l] Occurrences: 5,983 Examples: * cages: [c eː dʒ ɪ z] * jerry: [dʒ ɛ ɾ i] * surge: [s ɜː dʒ] * gpus: [dʒ iː p iː j ʉː z] |
||||||||
Sibilant |
Occurrences: 34,586 Examples: * stand: [s ʈ a n ɖ] * scots: [s k ɒ ʈ s] * mixed: [mʲ ɪ k s] * tulsa: [ʈ ə l s ə] Occurrences: 14,327 Examples: * loans: [l oː n z] * sings: [s ɪ n z] * minas: [mʲ ɪ n ə z] * goers: [ɡ ɜː ɹ ə z] |
Occurrences: 6,851 Examples: * shuts: [ʃ ə ʈ s] * shows: [ʃ oː dʒ] * shute: [ʃ ʉː ʈ] * blush: [b l ə ʃ] Occurrences: 500 Examples: * azure: [a ʒ ʊ ə] * anjou: [ɒ n ʒ ʉː] * loge: [l oː ʒ] * luger: [l ʉː ʒ ə] |
|||||||
Fricative |
Occurrences: 7,905 Examples: * find: [f aj n ɖ] * fake: [f eː k] * wifi: [ʋ aj f aj] * wilf: [ʋ ɪ l f] Occurrences: 2,306 Examples: * films: [fʲ ɪ l m dʒ] * fits: [fʲ ɪ ʈ s] * filly: [fʲ ɪ ʎ i] * fecal: [fʲ iː k ə l] |
Occurrences: 984 Examples: * hero: [ç ɪ ɾ oː] * haley: [ç eː ʎ i] * heed: [ç iː ɖ] * hale: [ç eː l] |
Occurrences: 3,762 Examples: * horst: [h ɒː s ʈ] * harpy: [h ɑː pʲ i] * hopi: [h oː pʲ i] * herne: [h ɜː n] |
||||||
Approximant |
Occurrences: 10 Examples: |
Occurrences: 11,160 Examples: * anwar: [a n ʋ ə] * luv: [l ə ʋ] * avon: [eː ʋ ɑ n] * heavy: [h ɛ ʋ i] |
Occurrences: 25,685 Examples: * elora: [ɪ l ɒː ɹ ə] * jura: [dʒ ʊ ɹ ə] * brain: [b ɹ eː n] * era: [ɪ ə ɹ ə] |
Occurrences: 1,471 Examples: * using: [j ʉː dʒ ɪ ŋ] * yates: [j eː ʈ s] * yolk: [j oː k] * usa: [j ʉː ɛ s eː] |
|||||
Tap |
Occurrences: 5,300 Examples: * genre: [dʒ a n ɾ ə] * foyer: [f ɔj ə ɾ] * breed: [b ɾ iː ɖ] * tried: [ʈ ɾ aj ɖ] |
||||||||
Lateral |
Occurrences: 23,471 Examples: * model: [m ɒ ɖ ə l] * carl: [k ɑː l] * lope: [l oː p] * pill: [p ɪ l] |
Occurrences: 8,546 Examples: * alesi: [eː ʎ iː s i] * leila: [ʎ iː l ə] * lymph: [ʎ ɪ m f] * foley: [f oː ʎ i] |
Vowels#
Vowel symbols to the left of are unrounded and those to the right are rounded.
Front |
Near-Front |
Central |
Near-Back |
Back |
|
---|---|---|---|---|---|
Close |
Occurrences: 12,826 Examples: * moly: [m oː ʎ i] * leafy: [ʎ iː fʲ i] * paddy: [p a ɖ i] * foci: [f ɒ tʃ i] Occurrences: 9,354 Examples: * kiev: [c iː ɛ ʋ] * we'd: [ʋ iː ɖ] * bleak: [b ʎ iː k] * liam: [ʎ iː ə m] |
Occurrences: 846 Examples: * bruch: [b ɹ ʉ k] * chu: [tʃ ʉ] * matsu: [m a ʈ s ʉ] * ubu: [ʉ b ʉː] Occurrences: 6,033 Examples: * sue: [s ʉː] * boo: [b ʉː] * fuel: [fʲ ʉː l] * boom: [b ʉː m] |
|||
Occurrences: 45,459 Examples: * spits: [s p ɪ ʈ s] * ears: [ɪ ə dʒ] * lynde: [ʎ ɪ n ɖ] * mere: [mʲ ɪ ə] |
Occurrences: 3,084 Examples: * jung: [j ʊ ŋ] * room: [ɹ ʊ m] * sushi: [s ʊ ʃ i] * urdu: [ʊ ə ɖ ʉː] |
||||
Close-Mid |
Occurrences: 0 Examples: Occurrences: 10,720 Examples: * grave: [ɟ ɹ eː ʋ] * seiun: [s eː ə n] * ways: [ʋ eː z] * lays: [l eː z] |
Occurrences: 0 Examples: Occurrences: 9,975 Examples: * codex: [k oː ɖ ɛ k s] * ropes: [ɾ oː p s] * hoe: [h oː] * rome: [ɹ oː m] |
|||
Occurrences: 64,935 Examples: * near: [ɲ ɪ ə] * angel: [eː n dʒ ə l] * cure: [c ʊ ə] * hazen: [ç eː z ə n] |
|||||
Open-Mid |
Occurrences: 16,279 Examples: * jenny: [dʒ ɛ ɲ i] * kiev: [c iː ɛ f] * quest: [cʷ ɛ s] * devil: [ɖ ɛ ʋ ə l] Occurrences: 1,123 Examples: * bears: [b ɛː z] * faria: [f ɛː ɹ ɪ ə] * airy: [ɛː ɹ i] * aram: [ɛː ɹ ə m] |
Occurrences: 191 Examples: * turku: [ʈ ɜ k ʉː] * circe: [s ɜ s i] * murr: [m ɜ] * blur: [b l ɜ] Occurrences: 3,335 Examples: * term: [ʈ ɜː m] * verb: [ʋ ɜː b] * percy: [p ɜː s i] * pasir: [p ə s ɜː] |
|||
Open |
Occurrences: 15,182 Examples: * drop: [dʒ ɹ a p] * matte: [m a ʈ] * anne: [a n] * slam: [s l a m] Occurrences: 519 Examples: * baba: [b aː b ə] * smart: [s m aː ʈ] * ranch: [ɾ aː n tʃ] * march: [m aː tʃ] |
Occurrences: 736 Examples: * arbor: [ɑ b ə] * makro: [m ɑ k ɹ oː] * huang: [ʋ ɑ ŋ] * saco: [s ɑ k oː] Occurrences: 5,654 Examples: * armey: [ɑː mʲ i] * pass: [p ɑː s] * bari: [b ɑː ɹ i] * nat: [n ɑː ʈ] Occurrences: 10,497 Examples: * quan: [kʷ ɒ n] * combo: [k ɒ m b oː] * chios: [c iː ɒ s] * keogh: [c iː ɒ ɡ] Occurrences: 5,211 Examples: * torch: [ʈ ɒː tʃ] * force: [f ɒː s] * corps: [k ɒː dʒ] * hoorn: [h ʊ ɒː n] |
Diphthongs#
aj
aw
ɔj