English (Nonnative) MFA dictionary v3.1.0#
@techreport{mfa_english_nonnative_mfa_dictionary_2024,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={English (Nonnative) MFA dictionary v3.1.0},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/English/English (Nonnative) MFA dictionary v3_1_0.html}},
year={2024},
month={Jun},
}
|
Installation#
Install from the MFA command line:
mfa model download dictionary english_nonnative_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 (Nonnative) MFA dictionary v3_1_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: 6,025 Examples: * mile: [m aj l] * sum: [s ɐ m] * sham: [ʃ a m] * name: [n e m] Occurrences: 1,142 Examples: * mini: [mʲ i ɲ i] * mills: [mʲ ɪ l z] * semi: [s ɛ mʲ i] * amino: [a mʲ iː n o] Occurrences: 1 Examples: |
Occurrences: 3 Examples: |
Occurrences: 13,257 Examples: * nara: [n ɑː ɹ ə] * sandy: [s a n dʲ i] * inch: [i n tʃ] * monte: [m ɒ n tʲ i] |
Occurrences: 2,126 Examples: * nina: [ɲ iː n a] * nitty: [ɲ i tʲ i] * linen: [ʎ i ɲ ɛ n] * nicks: [ɲ i k s] |
Occurrences: 2,614 Examples: * yong: [j ɑ ŋ] * along: [a l ɔ ŋ ɡ] * link: [ʎ i ŋ k] * ping: [p ɪ ŋ] |
||||
Stop Plain |
Occurrences: 5,497 Examples: * pick: [p ɪ k] * clasp: [c ʎ æ s p] * jumps: [dʒ ɐ m p s] * caps: [cʰ æ p s] Occurrences: 206 Examples: * speak: [s pʲ i k] * bumpy: [b ɐ m pʲ i] * speed: [s pʲ iː d] * wasps: [ʋ ɒ s pʲ i ɛ s] Occurrences: 3 Examples: Occurrences: 3,764 Examples: * grabs: [ɡ ɹ a b z] * bound: [b aw n d] * borax: [b ɒ ɹ a k s] * boyd: [b ɔj d] Occurrences: 599 Examples: * biggs: [bʲ i ɡ z] * usb: [j ʉː ɛ s bʲ ɪ] * beet: [bʲ iː t] * pubic: [pʲ uː bʲ i k] |
Occurrences: 532 Examples: * smith: [s mʲ i t̪] * broth: [b ɹ ɒ t̪] * frith: [f ɹ ɪ t̪] * theo: [t̪ iː oː] Occurrences: 133 Examples: * then: [d̪ ə n] * lathe: [l eː d̪] * paths: [p ɑː d̪ s] * than: [d̪ a n] |
Occurrences: 6,558 Examples: * notes: [n əw t s] * vomit: [v ɑ mʲ ɪ t] * mint: [mʲ i n t] * adult: [a d ɐ ɫ t] Occurrences: 2,046 Examples: * stena: [s tʲ iː n ə] * tnt: [tʲ iː ɛ n tʲ iː] * steal: [s tʲ iː ɫ] * deity: [d e i tʲ i] Occurrences: 37 Examples: * tweet: [tʷ iː t] * twig: [tʷ ɪ ɡ] * twist: [tʷ i s] Occurrences: 6,108 Examples: * dummy: [d ɐ mʲ i] * finds: [f aj n d z] * lied: [l aj d] * fond: [f ɑ n d] Occurrences: 1,558 Examples: * dig: [dʲ i dʒ] * video: [v ɪ dʲ i əw] * bandy: [b a n dʲ i] * ideal: [aj dʲ i a ɹ] |
Occurrences: 3,849 Examples: * pit: [p ɪ ʈ] * taffy: [ʈ a fʲ i] * tec: [ʈ ɛ k] * table: [ʈ eː b ə ɹ] Occurrences: 745 Examples: * putin: [p ʉː ʈʲ ɪ n] * piety: [p aj ɪ ʈʲ i] * ponte: [p ɒ n ʈʲ i] * putty: [p ə ʈʲ i] Occurrences: 4 Examples: Occurrences: 1,423 Examples: * woody: [ʋ ʊ ɖ i] * world: [ʋ ɜː ɹ ɖ] * weird: [ʋ ɪ ə ɖ] * panda: [p a n ɖ ə] |
Occurrences: 2,035 Examples: * kids: [c i ɖ s] * keras: [c ɛ ɹ ə z] * kyoto: [c j oː ʈ oː] * pekin: [p eː c ɪ n] Occurrences: 139 Examples: * quint: [cʷ i n t] * queer: [cʷ i a] * quiz: [cʷ i z] * quill: [cʷ ɪ l] Occurrences: 499 Examples: * enugu: [ɛ ɲ ə ɟ ʉː] * ugly: [ɐ ɟ ɹ i] * gills: [ɟ ɪ l z] * grim: [ɟ ɹ i m] Occurrences: 29 Examples: * gwynn: [ɟʷ ɪ n] * gwen: [ɟʷ ɛ n] |
Occurrences: 6,831 Examples: * cards: [k ɑː ɖ z] * perks: [p ɑː k s] * erich: [ɛ ɹ ɪ k] * erik: [ɛ ɹ ɪ k] Occurrences: 161 Examples: * squat: [s kʷ ɑ t] * choir: [kʷ aj a] * quiet: [kʷ aj i t] * quasi: [kʷ ɑ s i] Occurrences: 2,457 Examples: * rug: [ɹ ɐ ɡ] * pig: [p ɪ ɡ] * goods: [ɡ ʊ d s] * gaby: [ɡ a bʲ i] Occurrences: 21 Examples: * guam: [ɡʷ a m] * magua: [m ɑ ɡʷ ɑ] * guo: [ɡʷ əw] * guano: [ɡʷ a n o] |
Occurrences: 3 Examples: * often: [ɒ f ʔ ə n] |
||
Aspirated |
Occurrences: 281 Examples: * epoxy: [ɛ pʰ ɑ k s i] * plc: [pʰ iː ɛ ɫ s iː] * appel: [a pʰ ɛ ɫ] * japan: [dʒ a pʰ a n] |
Occurrences: 539 Examples: * moto: [m əw tʰ əw] * atari: [a tʰ a ɹ i] * mateo: [m a tʰ e o] * cacti: [cʰ æ k tʰ aj] |
Occurrences: 652 Examples: * kylie: [cʰ aj ʎ i] * camel: [cʰ æ m ə ɫ] * kyle: [cʰ aj ɫ] * camps: [cʰ æ m p s] |
Occurrences: 261 Examples: * cours: [kʰ ɒ ɹ z] * decoy: [d ɛ kʰ ɔj] * loco: [l əw kʰ əw] * coote: [kʰ ʉː t] |
|||||
Affricate |
Occurrences: 1,285 Examples: * chew: [tʃ i ʊ] * chola: [tʃ ow l ə] * trust: [tʃ ɹ ɐ s t] * hitch: [ç i tʃ] Occurrences: 1,819 Examples: * jays: [dʒ e z] * kenji: [c ɛ n dʒ i] * jury: [dʒ u ɹ i] * jeter: [dʒ i t ɚ] |
||||||||
Sibilant |
Occurrences: 13,725 Examples: * also: [ɒ ɹ s ow] * aunts: [ɑː n t s] * stiff: [s tʲ i f] * small: [s m ɒ ɫ] Occurrences: 6,069 Examples: * bases: [b e s iː z] * sells: [s ɛ l z] * vines: [v aj n z] * forms: [f ɒ ɹ m z] |
Occurrences: 2,725 Examples: * flash: [f l a ʃ] * she's: [ʃ ɛ z] * chico: [ʃ iː k ow] * vinci: [v ɪ n ʃ i] Occurrences: 95 Examples: * azul: [a ʒ ɒː l] * yonge: [j ɒ n ʒ] |
|||||||
Fricative |
Occurrences: 2,851 Examples: * fires: [f aj ə z] * serif: [s ɛ ɹ a f] * flag: [f l a ɡ] * faces: [f e s ɛ s] Occurrences: 753 Examples: * feces: [fʲ iː s iː z] * film: [fʲ ɪ l m] * feeds: [fʲ iː d z] * fees: [fʲ iː z] Occurrences: 2,134 Examples: * rive: [ɹ aj v] * leave: [ʎ i v] * vein: [v e n] * hove: [h əw v] Occurrences: 149 Examples: * gravy: [ɡ ɹ e vʲ i] * dvd: [dʲ iː vʲ iː dʲ iː] * vin: [vʲ ɪ n] * davey: [d e vʲ i] |
Occurrences: 208 Examples: * faith: [f e θ] * firth: [f aː θ] * cathy: [cʰ æ θ i] * mirth: [m aː θ] Occurrences: 123 Examples: * those: [ð əw z] * oaths: [əw ð z] * the: [ð a] * dhabi: [ð a bʲ i] |
Occurrences: 372 Examples: * hero: [ç ɪ ɹ əw] * humor: [ç uː m ɔ] * whom: [ç ʉː m] * hecla: [ç ɪ k l ə] |
Occurrences: 1,325 Examples: * mahan: [m ə h ɑː n] * harms: [h ɑ ɹ m z] * rahul: [ɹ ɑː h ʉː l] * hai: [h aj] |
|||||
Approximant |
Occurrences: 309 Examples: * irwin: [ɝ w ɪ n] * away: [a w e] * dwell: [d w ɛ ɫ] * nowak: [n əw w a k] |
Occurrences: 1,926 Examples: * wilma: [ʋ ɪ l m ə] * woods: [ʋ ʊ ɖ z] * walt: [ʋ ɒː l ʈ] * one's: [ʋ ə n s] |
Occurrences: 12,188 Examples: * learn: [ɹ aː n] * rank: [ɹ a ŋ k] * probe: [p ɹ əw b] * fries: [f ɹ aj s] |
Occurrences: 461 Examples: * used: [j uː s] * yoko: [j əw kʰ əw] * usual: [j uː ʃ] * yun: [j ɐ n] |
|||||
Tap |
Occurrences: 122 Examples: |
||||||||
Lateral |
Occurrences: 6,032 Examples: * lobos: [l ow b ow z] * lays: [l e z] * allan: [a l a n] * elvis: [ɛ l ʋ i s] Occurrences: 2,370 Examples: * small: [s m ɒ ɫ] * able: [b ə ɫ] * adult: [a d ɐ ɫ t] * label: [l e b ɛ ɫ] |
Occurrences: 2,887 Examples: * willi: [w ɪ ʎ i] * pliny: [p ʎ i ɲ i] * lind: [ʎ i n d] * lisp: [ʎ i s p] |
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: 9,754 Examples: * lacey: [l eː s i] * ernie: [ɝ ɲ i] * icy: [aj s i] * she: [ʃ i] Occurrences: 2,195 Examples: * neo: [ɲ iː əw] * seek: [s iː k] * suite: [s ʋ iː ʈ] * hemel: [ç iː m ə l] |
Occurrences: 418 Examples: * proof: [p ɹ ʉ f] * jewel: [dʒ ʉ ə ɫ] * gules: [ɡ ʉ ɫ z] * plume: [p l ʉ m] Occurrences: 1,065 Examples: * liu: [ʎ i ʉː] * shui: [ʃ ʉː i] * situ: [s ɪ ʈ ʉː] * poole: [p ʉː l] |
Occurrences: 358 Examples: * broom: [b ɹ u m] * hutu: [h u tʰ uː] * jason: [dʒ e s u n] * mason: [m e s u n] Occurrences: 630 Examples: * roost: [ɹ uː s t] * fruit: [f ɹ uː] * uni: [j uː ɲ iː] * loot: [l uː t] |
||
Occurrences: 10,310 Examples: * poems: [p oː ɪ m z] * renee: [ɹ ɪ ɲ eː] * tim: [ʈ ɪ m] * fiske: [fʲ ɪ s k] |
Occurrences: 854 Examples: * fully: [f ʊ ʎ i] * crook: [k ɹ ʊ k] * murad: [m ʊ ɹ ɑ d] * puna: [p ʊ n ɑː] |
||||
Close-Mid |
Occurrences: 1,753 Examples: * spade: [s p e d] * yale: [j e ɹ] * quake: [kʷ e k] * mane: [m e n] Occurrences: 718 Examples: * dea: [dʲ iː iː ej] * cafes: [cʰ æ f ej z] * cates: [cʰ ej t s] * mau: [ɛ m ej j ʉː] Occurrences: 1,540 Examples: * caves: [c eː ʋ s] * tape: [ʈ eː p] * che: [ʃ eː] * pekin: [p eː c ɪ n] |
Occurrences: 206 Examples: * idaho: [aj d a h o] * amino: [a mʲ iː n o] * odo: [ɔ d o] * samoa: [s a m o a] Occurrences: 307 Examples: * julio: [ç ʉː ʎ i ow] * hosea: [h ow z ej ə] * caddo: [cʰ æ d ow] * ufo: [j ʉ ɛ f ow] Occurrences: 681 Examples: * cody: [k oː ɖ i] * polo: [p oː ɹ oː] * tempo: [ʈ ɛ m p oː] * carlo: [k ɑː l oː] |
|||
Occurrences: 14,543 Examples: * dora: [d ɒː ɹ ə] * pulse: [p ə l s] * were: [ʋ ə] * yea: [j a ə] Occurrences: 744 Examples: * user: [j ʉ z ɚ] * ruler: [ɹ ʉ ɹ ɚ] * roger: [ɹ ɑ dʒ ɚ] * otter: [ɑ t ɚ] |
|||||
Open-Mid |
Occurrences: 7,688 Examples: * repel: [ɹ ɪ p ɛ l] * rouen: [ɹ ʉ ɛ n] * petty: [p ɛ ʈʲ i] * sided: [s aj d ɛ d] Occurrences: 168 Examples: * care: [c ɛː] * ears: [ɛː z] * piero: [p ɪ ɛː ɹ oː] * aire: [ɛː] |
Occurrences: 37 Examples: * turku: [ʈ ɜ k ʉː] * blur: [b l ɜ] * turvy: [ʈ ɜ ʋ i] Occurrences: 477 Examples: * per: [p ɜː] * terms: [ʈ ɜː m s] * pearl: [p ɜː l] * term: [ʈ ɜː m] Occurrences: 290 Examples: * ursus: [ɝ s ə s] * burt: [b ɝ t] * serbs: [s ɝ b z] * ernst: [ɝ n s t] |
Occurrences: 1,828 Examples: * bosom: [b ʊ z ɔ m] * atom: [a t ɔ m] * jacob: [dʒ e k ɔ b] * adopt: [a d ɔ p t] |
||
Occurrences: 792 Examples: * canis: [cʰ æ ɲ ɪ s] * cram: [c ɹ æ m] * cathy: [cʰ æ θ i] * chand: [tʃ æ n d] |
Occurrences: 2,051 Examples: * adult: [a d ɐ ɫ t] * until: [ɐ n tʲ i ɹ] * crust: [k ɹ ɐ s t] |
||||
Open |
Occurrences: 9,042 Examples: * slats: [s l a t s] * aloud: [a l aw d] * fiat: [f aj a t] * nang: [n a ŋ] Occurrences: 527 Examples: * merv: [m aː v] * serf: [s aː f] * bert: [b aː t] * nerd: [n aː d] |
Occurrences: 1,877 Examples: * lott: [l ɑ t] * hahn: [h ɑ n] * rohan: [ɹ əw h ɑ n] * jon: [dʒ ɑ n] Occurrences: 832 Examples: * kami: [k ɑː mʲ i] * card: [k ɑː ɖ] * parc: [p ɑː k] * ari: [ɑː ɹ i] Occurrences: 3,222 Examples: * wanda: [ʋ ɒ n ɖ ə] * molde: [m ɒ ɫ d] * wolof: [ʋ ɒ l ɒ f] * raw: [ɹ ɒ] Occurrences: 1,097 Examples: * wall: [ʋ ɒː ɹ] * paul: [p ɒː ɹ] * draws: [d ɹ ɒː s] * flaws: [f l ɒː z] |
Diphthongs#
aj
aw
ej
ow
ɔj
əw