English (US) MFA dictionary v2.0.0#
@techreport{mfa_english_us_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={English (US) MFA dictionary v2.0.0},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/English/English (US) MFA dictionary v2_0_0.html}},
year={2022},
month={Mar},
}
G2P models |
Installation#
Install from the MFA command line:
mfa model download dictionary english_us_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.
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: 17,463 Examples: * timer: [tʰ aj m ɚ] * mail: [m ej ɫ] * among: [ə m ɐ ŋ] * boom: [b ʉ m] Occurrences: 979 Examples: * mien: [mʲ iː n] * marmy: [m ɑ ɹ mʲ i] * rummy: [ɹ ɐ mʲ i] * gormy: [ɡ ɒ ɹ mʲ i] Occurrences: 1,511 Examples: * annum: [æ n m̩] * golem: [ɡ ow l m̩] * hmmm: [h m̩] * him: [h m̩] |
Occurrences: 1,090 Examples: * infer: [ɪ ɱ f ɝ] * comfy: [kʰ ɐ ɱ fʲ i] * envoy: [ɑ ɱ v ɔj] * senvy: [s ɛ ɱ vʲ i] |
Occurrences: 29,440 Examples: * yan: [j æ n] * yarns: [j ɑ ɹ n z] * downe: [d aw n] * towne: [tʰ aw n] Occurrences: 3,799 Examples: * aspen: [æ s p n̩] * semen: [s iː m n̩] * taken: [tʰ ej k n̩] * token: [tʰ ow k n̩] |
Occurrences: 7,059 Examples: * naiad: [ɲ ej æ d] * nim: [ɲ ɪ m] * nah: [ɲ æ] * genin: [dʒ ɛ ɲ ɪ n] |
Occurrences: 7,253 Examples: * among: [ə m ɐ ŋ] * pang: [pʰ æ ŋ] * crunk: [k ɹ ɐ ŋ k] * kang: [kʰ ɑ ŋ] |
|||
Stop |
Occurrences: 10,065 Examples: * stipe: [s t aj p] * tape: [tʰ ej p] * cup: [kʰ ɐ p] * provo: [p ɹ ow v ow] Occurrences: 13,001 Examples: * robed: [ɹ ow b d] * boom: [b ʉ m] * abase: [ə b ej s] * curb: [kʰ ɝ b] |
Occurrences: 23,170 Examples: * stipe: [s t aj p] * smut: [s m ɐ t] * clit: [c ʎ ɪ t] * eked: [iː k t] Occurrences: 18,747 Examples: * robed: [ɹ ow b d] * downe: [d aw n] * naiad: [n aj æ d] * radar: [ɹ ej d ɑ ɹ] |
Occurrences: 4,389 Examples: * clit: [c ʎ ɪ t] * leaky: [ʎ iː c i] * klang: [c ʎ æ ŋ] * crain: [c ɹ ej n] Occurrences: 2,931 Examples: * foggy: [f ɑ ɟ i] * ge'ez: [ɟ iː ɛ z] * glatt: [ɟ ʎ æ t] * digby: [d ɪ ɟ bʲ i] |
Occurrences: 14,457 Examples: * yerk: [j ɝ k] * eked: [iː k t] * krook: [k ɹ ʊ k] * obex: [ow b ɛ k s] Occurrences: 5,085 Examples: * gafia: [ɡ æ fʲ i ə] * gum: [ɡ ɐ m] * fargo: [f ɑ ɹ ɡ ow] * goats: [ɡ ow t s] |
Occurrences: 1,282 Examples: * smut: [s m ɐ ʔ] * visit: [v ɪ z ɪ ʔ] * hmmm: [ʔ m̩] * bert: [b ɝ ʔ] |
|||
Affricate |
Occurrences: 3,230 Examples: * oroch: [ɒː ɹ ɒ tʃ] * itch: [ɪ tʃ] * birch: [b ɝ tʃ] * chafe: [tʃ ej f] Occurrences: 4,938 Examples: * jorge: [dʒ ɒ ɹ dʒ] * podgy: [pʰ ɑ dʒ i] * gwas: [dʒ i w ɑ s] * joab: [dʒ ow æ b] |
|||||||
Sibilant |
Occurrences: 36,172 Examples: * sir: [s ɝ] * sophy: [s ow fʲ i] * stipe: [s t aj p] * abase: [ə b ej s] Occurrences: 15,465 Examples: * yarns: [j ɑ ɹ n z] * raze: [ɹ ej z] * visit: [v ɪ z ɪ t] * exist: [ɪ ɡ z ɪ s t] |
Occurrences: 6,752 Examples: * shea: [ʃ ej] * shmoo: [ʃ m ʉː] * shunt: [ʃ ɐ n t] * frush: [f ɹ ɐ ʃ] Occurrences: 551 Examples: * zhe: [ʒ iː] * liege: [ʎ iː ʒ] * uzhe: [j ʉː ʒ] * taj: [tʰ ɑ ʒ] |
||||||
Fricative |
Occurrences: 9,599 Examples: * fargo: [f ɑ ɹ ɡ ow] * infer: [ɪ ɱ f ɝ] * reefs: [ɹ iː f s] * foggy: [f ɑ ɟ i] Occurrences: 688 Examples: * sophy: [s ow fʲ i] * gafia: [ɡ æ fʲ i ə] * fees: [fʲ iː z] * fusil: [fʲ ʉː z ɪ ɫ] Occurrences: 6,907 Examples: * wolve: [w ʊ ɫ v] * visit: [v ɪ z ɪ t] * provo: [p ɹ ow v ow] * levee: [l ə v ej] Occurrences: 437 Examples: * levee: [l ə vʲ i] * livy: [ʎ ɪ vʲ i] * privy: [p ɹ ɪ vʲ i] * davie: [d ej vʲ i] |
Occurrences: 2,506 Examples: * thro: [θ ɹ ow] * thru: [θ ɹ ʉ] * thick: [θ ɪ k] * saith: [s ej ə θ] Occurrences: 633 Examples: * other: [ɐ ð ɚ] * with: [w ɪ ð] * dhikr: [ð i k ɹ] * theys: [ð ej z] |
Occurrences: 924 Examples: * jorge: [h ɒ ɹ ç ej] * him: [ç ɪ m] * hayat: [ç ej ɑ t] * huron: [ç ɝ ɑ n] |
Occurrences: 3,736 Examples: * jorge: [h ɒ ɹ ç ej] * toho: [tʰ ow h ow] * hal: [h æ ɫ] * hmmm: [h m̩] |
||||
Approximant |
Occurrences: 5,661 Examples: * wolve: [w ʊ ɫ v] * iwar: [aj w ɒ ɹ] * kauai: [kʰ ə w aj iː] * gwas: [dʒ i w ɑ s] |
Occurrences: 35,096 Examples: * robed: [ɹ ow b d] * jorge: [dʒ ɒ ɹ dʒ] * awry: [ə ɹ aj] * yarns: [j ɑ ɹ n z] |
Occurrences: 1,188 Examples: * yan: [j æ n] * yarns: [j ɑ ɹ n z] * yerk: [j ɝ k] * yeast: [j iː s t] |
|||||
Tap |
Occurrences: 10,275 Examples: * radar: [ɹ ej ɾ ɚ] * maddy: [m æ ɾ i] * part: [pʰ ɑ ɹ ɾ] * cital: [s aj ɾ ɫ̩] |
|||||||
Lateral |
Occurrences: 12,159 Examples: * leila: [l aj l ə] * glegg: [ɡ l ɛ ɡ] * sola: [s ow l ə] * levee: [l ə v ej] Occurrences: 6,181 Examples: * mail: [m ej ɫ] * towel: [tʰ aw ɫ] * wolve: [w ʊ ɫ v] * guirl: [ɡ ɝ ɫ] Occurrences: 4,993 Examples: * towel: [tʰ aw ɫ̩] * cital: [s aj ɾ ɫ̩] * bagel: [b ej ɡ ɫ̩] * kabul: [kʰ ɑː b ɫ̩] |
Occurrences: 9,688 Examples: * telly: [tʰ ɛ ʎ i] * leila: [ʎ iː l ə] * clit: [c ʎ ɪ t] * leaky: [ʎ iː c 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: 16,484 Examples: * sophy: [s ow fʲ i] * telly: [tʰ ɛ ʎ i] * podgy: [pʰ ɑ dʒ i] * gafia: [ɡ æ fʲ i ə] Occurrences: 6,275 Examples: * leila: [ʎ iː l ə] * eked: [iː k t] * mien: [mʲ iː n] * kauai: [kʰ ə w aj iː] |
Occurrences: 2,353 Examples: * boom: [b ʉ m] * lewd: [ʎ ʉ d] * thru: [θ ɹ ʉ] * kunes: [cʰ ʉ n z] Occurrences: 4,564 Examples: * goops: [ɡ ʉː p s] * moods: [m ʉː d z] * grew: [ɡ ɹ ʉː] * shmoo: [ʃ m ʉː] |
|||
Occurrences: 43,918 Examples: * clit: [c ʎ ɪ t] * visit: [v ɪ z ɪ t] * shea: [ʃ ɪ ə] * nim: [ɲ ɪ m] |
Occurrences: 2,532 Examples: * wolve: [w ʊ ɫ v] * bourn: [b ʊ ɹ n] * krook: [k ɹ ʊ k] * stood: [s t ʊ d] |
||||
Close-Mid |
Occurrences: 11,362 Examples: * epee: [ej pʰ ej] * mail: [m ej ɫ] * jorge: [h ɒ ɹ ç ej] * abase: [ə b ej s] |
Occurrences: 10,115 Examples: * robed: [ɹ ow b d] * sophy: [s ow fʲ i] * toho: [tʰ ow h ow] * fargo: [f ɑ ɹ ɡ ow] |
|||
Occurrences: 36,789 Examples: * among: [ə m ɐ ŋ] * awry: [ə ɹ aj] * abase: [ə b ej s] * leila: [l aj l ə] Occurrences: 9,416 Examples: * timer: [tʰ aj m ɚ] * radar: [ɹ ej ɾ ɚ] * borer: [b ɒː ɹ ɚ] * tamer: [tʰ ej m ɚ] |
|||||
Open-Mid |
Occurrences: 18,685 Examples: * epee: [ɛ pʰ ej] * yan: [j ɛ n] * telly: [tʰ ɛ ʎ i] * glegg: [ɡ l ɛ ɡ] |
Occurrences: 3,470 Examples: * sir: [s ɝ] * yerk: [j ɝ k] * curb: [kʰ ɝ b] * infer: [ɪ ɱ f ɝ] |
|||
Occurrences: 15,937 Examples: * yan: [j æ n] * pang: [pʰ æ ŋ] * gafia: [ɡ æ fʲ i ə] * naiad: [n aj æ d] |
Occurrences: 7,321 Examples: * among: [ə m ɐ ŋ] * smut: [s m ɐ t] * gum: [ɡ ɐ m] * cup: [kʰ ɐ p] |
||||
Open |
Occurrences: 10,815 Examples: * yarns: [j ɑ ɹ n z] * podgy: [pʰ ɑ dʒ i] * gafia: [ɡ ɑ fʲ i ə] * fargo: [f ɑ ɹ ɡ ow] Occurrences: 1,919 Examples: * calas: [kʰ ə l ɑː] * nah: [n ɑː] * mach: [m ɑː k] * kabul: [kʰ ɑː b ɫ̩] Occurrences: 5,574 Examples: * jorge: [dʒ ɒ ɹ dʒ] * iwar: [aj w ɒ ɹ] * bourn: [b ɒ ɹ n] * oroch: [ɒː ɹ ɒ tʃ] Occurrences: 1,369 Examples: * oroch: [ɒː ɹ ɒ tʃ] * borer: [b ɒː ɹ ɚ] * paulo: [pʰ ɒː l ow] * tau: [tʰ ɒː] |
Diphthongs#
aj
aw
ej
ow
ɔj