English (UK) MFA dictionary v2.2.1#
@techreport{mfa_english_uk_mfa_dictionary_2023,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={English (UK) MFA dictionary v2.2.1},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/English/English (UK) MFA dictionary v2_2_1.html}},
year={2023},
month={May},
}
G2P models |
Installation#
Install from the MFA command line:
mfa model download dictionary english_uk_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 (UK) MFA dictionary v2_2_1.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 |
Palatal |
Velar |
Glottal |
---|---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 15,537 Examples: * stum: [s t ɐ m] * proem: [p ɹ əw ə m] * matty: [m a tʲ i] * mohur: [m əw h ə] Occurrences: 3,401 Examples: * romeo: [ɹ əw mʲ ɪ əw] * mitt: [mʲ ɪ t] * ming: [mʲ ɪ ŋ] * barmy: [b ɑː mʲ i] Occurrences: 3 Examples: |
Occurrences: 1 Examples: |
Occurrences: 31,578 Examples: * brine: [b ɹ aj n] * unto: [ɐ n tʰ ʊ] * snart: [s n ɑː t] * ripon: [ɹ ɪ pʰ ɒ n] Occurrences: 166 Examples: |
Occurrences: 5,942 Examples: * fungi: [f ɐ ɲ ɟ iː] * nake: [ɲ ej k] * muni: [mʲ ʉː ɲ i] * nero: [ɲ iː ɹ əw] |
Occurrences: 5,983 Examples: * chang: [tʃ a ŋ] * jang: [dʒ a ŋ] * crunk: [k ɹ ɐ ŋ k] * aging: [ej dʒ ɪ ŋ] |
|||
Stop |
Occurrences: 8,959 Examples: * plow: [p l aw] * nopal: [n əw p ə ɫ] * helps: [h ɛ ɫ p s] * clap: [c ʎ a p] Occurrences: 10,188 Examples: * jibs: [dʒ ɪ b z] * barb: [b ɑː b] * urban: [ɜː b ə n] * bunny: [b ɐ ɲ i] |
Occurrences: 69 Examples: Occurrences: 66 Examples: |
Occurrences: 20,007 Examples: * juts: [dʒ ɐ t s] * storm: [s t ɒː m] * pesto: [pʰ ɛ s t əw] * utter: [ɐ t ə] Occurrences: 15,499 Examples: * qaeda: [cʰ ʉː ej d ə] * hamid: [h a mʲ ɪ d] * domal: [d əw m ə ɫ] * breda: [b ɹ ej d ɑː] |
Occurrences: 3,790 Examples: * wiki: [w ɪ c i] * skene: [s c iː ɲ i] * porky: [pʰ ɒː c i] * pinky: [pʰ ɪ ɲ c i] Occurrences: 1,666 Examples: * algae: [a ɫ ɟ i] * get: [ɟ ɪ t] * glean: [ɟ ʎ iː n] * gyoza: [ɟ əw z ə] |
Occurrences: 13,330 Examples: * exon: [ɛ k s ə n] * atka: [ɑː t k ə] * dafuq: [d ə f ɐ k] * fatx: [f a tʰ ɛ k s] Occurrences: 5,687 Examples: * bugan: [b ɐ ɡ ə n] * gowns: [ɡ aw n z] * gabon: [ɡ ə b ɒ n] * gup: [ɡ ɐ p] |
Occurrences: 482 Examples: * spot: [s p ɒ ʔ] * hat: [h a ʔ] * item: [aj ʔ ə m] * flat: [f l a ʔ] |
||
Affricate |
Occurrences: 2,884 Examples: * dutch: [d ɐ tʃ] * cho: [tʃ əw] * tree: [tʃ iː] * botch: [b ɒ tʃ] Occurrences: 4,803 Examples: * jima: [dʒ ɪ m ə] * dpg: [dʲ iː pʰ iː dʒ iː] * gmpp: [dʒ iː ɛ m pʰ iː pʰ iː] * jeter: [dʒ iː t ə] |
|||||||
Sibilant |
Occurrences: 32,893 Examples: * syn: [s ɪ n] * son: [s ɒ n] * sata: [s ej t ə] * json: [dʒ ej s ə n] Occurrences: 13,431 Examples: * crows: [k ɹ əw z] * noisy: [n ɔj z i] * aegis: [iː dʒ ɪ dʲ iː z] * gobs: [ɡ ɒ b z] |
Occurrences: 6,485 Examples: * shish: [ʃ ɪ ʃ] * ghosh: [ɡ əw ʃ] * apish: [ej p ɪ ʃ] * shad: [ʃ a d] Occurrences: 496 Examples: * ezhes: [ɛ ʒ ɪ z] * zhe: [ʒ iː] * zhuzh: [ʒ ʊ ʒ] * azul: [a ʒ ɒː ɫ] |
||||||
Fricative |
Occurrences: 7,323 Examples: * fock: [f ɒ k] * smurf: [s m ɜː f] * morph: [m ɒː f] * skift: [s c ɪ f t] Occurrences: 2,184 Examples: * fuchs: [fʲ ʉː k s] * cafe: [kʰ a fʲ i] * fios: [fʲ ɪ s] * beefy: [bʲ iː fʲ i] Occurrences: 3 Examples: * fwoom: [fʷ ʉː m] Occurrences: 6,134 Examples: * veils: [v ej ɫ z] * rival: [ɹ aj v ə ɫ] * have: [h a v] * voted: [v əw tʲ ɪ d] Occurrences: 546 Examples: * mvp: [ɛ m vʲ iː pʰ iː] * vila: [vʲ iː l ə] * mtv: [ɛ m tʰ iː vʲ iː] * cofe: [s iː ə vʲ iː] Occurrences: 8 Examples: * vworp: [vʷ ɒː p] |
Occurrences: 2,359 Examples: * thor: [θ ɒː] * thoth: [θ əw θ] * athos: [a θ ɒ s] * dryth: [d ɹ aj θ] Occurrences: 525 Examples: * lithe: [l aj ð] * with: [w ɪ ð] * those: [ð əw z] * oaths: [əw ð z] |
Occurrences: 918 Examples: * hyp: [ç ɪ p] * heal: [ç iː ɫ] * hiply: [ç ɪ p ʎ i] * hits: [ç ɪ t s] |
Occurrences: 3,523 Examples: * hope: [h əw p] * haut: [h ɒː t] * behan: [b ə h ɑː n] * hells: [h ɛ ɫ z] |
||||
Approximant |
Occurrences: 3,790 Examples: * noir: [n w ɑː] * swalk: [s w ɒː ɫ k] * wiley: [w aj ʎ i] * woy: [w ɔj] |
Occurrences: 25,101 Examples: * ruga: [ɹ ʉː ɡ ə] * sari: [s ɑː ɹ i] * eric: [ɛ ɹ ɪ k] * hijra: [ç ɪ dʒ ɹ ə] |
Occurrences: 1,367 Examples: * yup: [j ɐ p] * yelps: [j ɛ ɫ p s] * ukase: [j ʉː cʰ ej z] * utica: [j ʉː tʲ ɪ k ə] |
|||||
Lateral |
Occurrences: 11,706 Examples: * luo: [l ʉː əw] * lesbo: [l ɛ z b əw] * lyne: [l aj n] * alary: [ej l ə ɹ i] Occurrences: 10,371 Examples: * diels: [d aj ə ɫ z] * mohel: [m ɔj ə ɫ] * elcic: [iː ɛ ɫ s iː aj s iː] * smile: [s m aj ə ɫ] Occurrences: 142 Examples: |
Occurrences: 8,151 Examples: * hilly: [ç ɪ ʎ i] * gully: [ɡ ɐ ʎ i] * limp: [ʎ ɪ m p] * cilia: [s ɪ ʎ i ə] |
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: 12,244 Examples: * anti: [a n tʲ i] * pory: [pʰ ɒː ɹ i] * filmy: [fʲ ɪ ɫ mʲ i] * loewy: [l ʉː i] Occurrences: 8,874 Examples: * vth: [vʲ iː tʰ iː ej tʃ] * pmee: [pʰ iː mʲ iː] * plea: [p ʎ iː] * kiwis: [cʰ iː w i z] |
Occurrences: 816 Examples: * lujvo: [l ʉ ʒ v əw] * uluru: [ʉː l ʉ ɹ ʉː] * umar: [ʉ m ɑː] * bellu: [b ɛ ʎ ʉ] Occurrences: 5,753 Examples: * vroom: [v ɹ ʉː m] * used: [j ʉː z d] * rube: [ɹ ʉː b] * goto: [ɡ əw tʰ ʉː] |
|||
Occurrences: 42,620 Examples: * jilt: [dʒ ɪ ɫ t] * inman: [ɪ n m ə n] * evade: [ɪ v ej d] * neary: [ɲ ɪ ə ɹ i] |
Occurrences: 2,937 Examples: * shula: [ʃ ʊ l ə] * quran: [kʰ ʊ ə n] * unken: [ʊ ŋ k ə n] * kure: [kʰ ʊ ə] |
||||
Close-Mid |
Occurrences: 1 Examples: Occurrences: 10,077 Examples: * kate: [cʰ ej t] * aging: [ej dʒ ɪ ŋ] * fady: [f ej dʲ i] * gays: [ɡ ej z] |
||||
Occurrences: 52,861 Examples: * basta: [b a s t ə] * who: [d ɐ b ə ʎ ʉː ej tʃ əw] * cowra: [kʰ aw ɹ ə] * atom: [a t ə m] |
|||||
Open-Mid |
Occurrences: 15,191 Examples: * never: [n ɛ v ə] * ident: [aj d ɛ n t] * sql: [ɛ s c ʉː ɛ ɫ] * efta: [ɛ f tʰ ɑː] Occurrences: 1,047 Examples: * arian: [ɛː ɹ i ə n] * khene: [cʰ ɛː n] * share: [ʃ ɛː] * hare: [h ɛː] |
Occurrences: 190 Examples: * furor: [fʲ ɜ ɹ ə] * wirth: [w ɜ θ] * herzl: [h ɜ t s ə ɫ] * turku: [tʰ ɜ k ʉː] Occurrences: 3,196 Examples: * merle: [m ɜː ɫ] * verst: [v ɜː s t] * birsy: [b ɜː s i] * curio: [cʰ ɜː ɹ i əw] |
|||
Occurrences: 6,301 Examples: * govvy: [ɡ ɐ vʲ i] * sough: [s ɐ f] * nandi: [n ɐ n dʲ i] * sufi: [s ɐ fʲ i] |
|||||
Open |
Occurrences: 14,077 Examples: * ratt: [ɹ a t] * knap: [n a p] * badly: [b a dʲ ʎ i] * imax: [aj m a k s] |
Occurrences: 737 Examples: * rohan: [ɹ əw h ɑ n] * vann: [v ɑ n] * waag: [w ɑ ɡ] * baca: [b ə kʰ ɑ] Occurrences: 5,483 Examples: * maar: [m ɑː] * douar: [d ʉː ɑː] * varna: [v ɑː n ə] * nazir: [n ɑː z ɪ ə] Occurrences: 10,048 Examples: * oka: [ɒ k ə] * onto: [ɒ n tʰ ʉː] * boron: [b ɒː ɹ ɒ n] * coxa: [kʰ ɒ k s ə] Occurrences: 4,935 Examples: * warns: [w ɒː n z] * malt: [m ɒː ɫ t] * mhorr: [m ɒː] * vworp: [vʷ ɒː p] |
Diphthongs#
aj
aw
ej
ɔj
əw