Ctrl+K
Ctrl+K

English (US) MFA dictionary v2.0.0a#

  • Maintainer: Montreal Forced Aligner

  • Language: English

  • Dialect: General American English

  • Phone set: MFA

  • Number of words: 79,080

  • Phones: aj aw b bʲ c cʰ d dʒ dʲ d̪ ej f fʲ h i iː j k kʰ l m mʲ m̩ n n̩ ow p pʰ pʲ s t tʃ tʰ tʲ t̪ v vʲ w z æ ç ð ŋ ɐ ɑ ɑː ɒ ɒː ɔj ə ɚ ɛ ɝ ɟ ɡ ɪ ɫ ɫ̩ ɱ ɲ ɹ ɾ ɾʲ ʃ ʉ ʉː ʊ ʎ ʒ ʔ θ

  • License: CC BY 4.0

  • Compatible MFA version: v2.0.0

  • Citation:

@techreport{mfa_english_us_mfa_dictionary_2022,
	author={McAuliffe, Michael and Sonderegger, Morgan},
	title={English (US) MFA dictionary v2.0.0a},
	address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/English/English (US) MFA dictionary v2_0_0a.html}},
	year={2022},
	month={May},
}

G2P models
../../_images/full_logo_yellow.svg

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:
17,131
Examples:
* demos:
[d ɛ m o w z]
* moore:
[m ʊ ɹ]
* sami:
[s ɑ ː m ʲ i]
* edm:
[ɛ d m]
Occurrences:
3,970
Examples:
* grimy:
[ɡ ɹ a j m ʲ i]
* media:
[m ʲ i ː d ʲ i ə]
* roomy:
[ɹ ʉ ː m ʲ i]
* amine:
[æ m ʲ i ː n]
Occurrences:
1,753
Examples:
* mgm:
[m ʲ ɪ ɡ m ̩]
* maxim:
[m æ k s m ̩]
* totem:
[t ʰ o w t m ̩]
* no'm:
[n o w m ̩]
Occurrences:
1,295
Examples:
* infra:
[ɪ ɱ f ɹ ə]
* mvp:
[ɛ ɱ v ʲ i ː p ʰ i ː]
* infix:
[ɪ ɱ f ʲ ɪ k s]
* banff:
[b æ ɱ f]
Occurrences:
33,097
Examples:
* indus:
[ɪ n d ə s]
* not:
[n ɑ t]
* nadia:
[n ɑ d ʲ i ə]
* airan:
[a j ɹ ɑ n]
Occurrences:
7,202
Examples:
* karan:
[c ʰ ɛ ɹ n ̩]
* boron:
[b ɒ ɹ n ̩]
* pagan:
[p ʰ e j ɡ n ̩]
* asian:
[e j ʒ n ̩]
Occurrences:
8,373
Examples:
* lenin:
[l ɛ ɲ ɪ n]
* manes:
[m ɑ ː ɲ e j z]
* lenny:
[l ɛ ɲ i]
* ennis:
[ɛ ɲ ɪ s]
Occurrences:
8,452
Examples:
* sung:
[s ɐ ŋ]
* ganga:
[ɡ ɑ ː ŋ ɡ ə]
* sank:
[s æ ŋ k]
* plunk:
[p l ɐ ŋ k]

Stop
Occurrences:
10,066
Examples:
* bacup:
[b e j k ʰ ɐ p]
* pario:
[p ʰ ɛ ɹ i o w]
* prill:
[p ɹ ɪ ɫ]
* paren:
[p ʰ ə ɹ ɛ n]
Occurrences:
12,687
Examples:
* barff:
[b ɑ ɹ f]
* rabe:
[ɹ e j b]
* byte:
[b a j t]
* brink:
[b ɹ ɪ ŋ k]
Occurrences:
99
Examples:
* thank:
[t ̪ æ ŋ k]
* think:
[t ̪ ɪ ŋ]
* both:
[b o w t ̪]
* thumb:
[t ̪ ɐ m]
Occurrences:
119
Examples:
* there:
[d ̪ ɚ]
* theys:
[d ̪ e j z]
* oaths:
[o w d ̪ z]
* that:
[d ̪ æ t]
Occurrences:
25,687
Examples:
* alate:
[e j l e j t]
* ftp:
[ɛ f t ʲ i ː p ʰ i ː]
* azote:
[æ z o w t]
* liter:
[ʎ i ː t ɚ]
Occurrences:
19,715
Examples:
* kited:
[c ʰ a j t ʲ ɪ d]
* jax:
[d ʒ æ k s]
* judy:
[d ʒ ʉ ː d ʲ i]
* dunks:
[d ɐ ŋ k s]
Occurrences:
5,764
Examples:
* skink:
[s c ɪ ŋ k]
* keke:
[c ʰ i ː k]
* katy:
[c ʰ e j t ʲ i]
* kilns:
[c ʰ ɪ ɫ n z]
Occurrences:
3,442
Examples:
* grier:
[ɟ ɹ ɪ ɹ]
* gish:
[ɟ ɪ ʃ]
* gink:
[ɟ ɪ ŋ k]
* guile:
[ɟ a j ɫ]
Occurrences:
15,737
Examples:
* trax:
[t ɹ æ k s]
* hawk:
[h ɒ k]
* jerk:
[d ʒ ɝ k]
* tcler:
[t ʰ ɪ k l ɚ]
Occurrences:
5,664
Examples:
* guts:
[ɡ ɐ t s]
* gusto:
[ɡ ɐ s t o w]
* gow:
[ɡ a w]
* dagan:
[d æ ɡ n ̩]
Occurrences:
1,719
Examples:
* greet:
[ɟ ɹ i ː ʔ]
* cut:
[k ʰ ɐ ʔ]
* skate:
[s c e j ʔ]
* yeet:
[j i ː ʔ]

Affricate
Occurrences:
3,501
Examples:
* chips:
[t ʃ ɪ p s]
* chill:
[t ʃ ɪ ɫ]
* situ:
[s i ː t ʃ ʉ]
* chap:
[t ʃ æ p]
Occurrences:
5,576
Examples:
* john:
[d ʒ ɑ n]
* jiggy:
[d ʒ ɪ ɟ i]
* joist:
[d ʒ ɔ j s t]
* mjg:
[ɛ m d ʒ e j d ʒ i ː]

Sibilant
Occurrences:
40,696
Examples:
* tops:
[t ʰ ɑ p s]
* slopy:
[s l o w p ʲ i]
* swink:
[s w ɪ ŋ k]
* strep:
[s t ɹ ɛ p]
Occurrences:
18,324
Examples:
* byers:
[b a j ɹ z]
* sizz:
[s ɪ z]
* ales:
[e j l z]
* gauze:
[ɡ ɒ ː z]
Occurrences:
9,770
Examples:
* sdch:
[s æ n d w ɪ t ʃ]
* apush:
[e j p ʰ ʊ ʃ]
* apish:
[e j p ʲ ɪ ʃ]
* ashes:
[æ ʃ ɪ z]
Occurrences:
744
Examples:
* jolt:
[d ʒ o w ɫ t]
* jawed:
[d ʒ ɒ ː d]
* gyrus:
[d ʒ a j ɹ ə s]
* gmpp:
[d ʒ i ː ɛ m p ʰ i ː p ʰ i ː]

Fricative
Occurrences:
9,087
Examples:
* faw:
[f ɒ ː]
* oft:
[ɒ ː f t]
* farre:
[f ɛ ɹ]
* flier:
[f l a j ɚ]
Occurrences:
2,640
Examples:
* flues:
[f ʲ ʎ ʉ ː z]
* finn:
[f ʲ ɪ n]
* flu:
[f ʲ ʎ ʉ ː]
* fiat:
[f ʲ i æ t]
Occurrences:
6,556
Examples:
* luvz:
[l ɐ v z]
* haven:
[ç e j v n ̩]
* aaav:
[e j e j e j v ʲ i ː]
* vis:
[v ʲ ɪ s]
Occurrences:
2,015
Examples:
* virid:
[v ʲ ɪ ɹ ɪ d]
* tavis:
[t ʰ æ v ʲ ɪ s]
* vigor:
[v ʲ ɪ ɡ ɚ]
* veal:
[v ʲ i ː ɫ]
Occurrences:
2,723
Examples:
* thaw:
[θ ɒ ː]
* thrum:
[θ ɹ ɐ m]
* mirth:
[m ɝ θ]
* thuja:
[θ ʉ d ʒ ə]
Occurrences:
615
Examples:
* th':
[ð]
* that:
[ð æ t]
* this:
[ð ɪ s]
* them:
[ð m ̩]
Occurrences:
1,080
Examples:
* herod:
[ç ɛ ɹ ə d]
* hit:
[ç ɪ t]
* hews:
[ç ʉ ː z]
* hadal:
[ç e j d ɫ ̩]
Occurrences:
4,130
Examples:
* lehua:
[l ɛ h ʉ ː ə]
* hot:
[h ɑ t]
* hardy:
[h ɑ ɹ d ʲ i]
* hurts:
[h ɝ t s]

Approximant
Occurrences:
6,425
Examples:
* odors:
[o w d ɚ z]
* wino:
[w ɪ n o w]
* tweak:
[t w i ː k]
* vogue:
[v o w ɡ]
Occurrences:
40,835
Examples:
* fakir:
[f ɑ c ɪ ɹ]
* realm:
[ɹ ɛ ɫ m]
* sears:
[s ɪ ɹ z]
* irony:
[a j ɹ ə ɲ i]
Occurrences:
1,286
Examples:
* mia:
[ɛ m a j e j]
* duvet:
[d ʲ ʉ ː v e j]
* piles:
[p ʰ a j ɫ z]
* baily:
[b e j ʎ i]

Tap
Occurrences:
4,860
Examples:
* vital:
[v a j ɾ ɫ ̩]
* anita:
[ə ɲ i ː ɾ ə]
* sotho:
[s o w ɾ o w]
* etc:
[ɛ k s ɛ ɾ ə ɹ ə]
Occurrences:
5,674
Examples:
* faded:
[f e j ɾ ʲ ɪ d]
* snati:
[s ɲ æ ɾ ʲ i]
* rorty:
[ɹ ɒ ɹ ɾ ʲ i]
* saudi:
[s a w ɾ ʲ i]

Lateral
Occurrences:
13,694
Examples:
* flex:
[f l ɛ k s]
* lye:
[l a j]
* luvs:
[l ɐ v z]
* lent:
[l ɛ n t]
Occurrences:
6,541
Examples:
* alma:
[æ ɫ m ə]
* patel:
[p ʰ ə t ʰ ɛ ɫ]
* spill:
[s p ʲ ɪ ɫ]
* wells:
[w ɛ ɫ z]
Occurrences:
5,799
Examples:
* sigil:
[s ɪ d ʒ ɫ ̩]
* hovel:
[h ɐ v ɫ ̩]
* vocal:
[v o w k ɫ ̩]
* nasal:
[ɲ e j z ɫ ̩]
Occurrences:
11,435
Examples:
* lethe:
[ʎ i ː θ i]
* lunar:
[ʎ ʉ n ɚ]
* delhi:
[d ɛ ʎ i]
* haply:
[h æ p ʲ 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:
19,731
Examples:
* ceil:
[s i ː ɫ]
* gulfy:
[ɡ ɐ ɫ f ʲ i]
* teem:
[t ʲ i ː m]
* curia:
[c ʰ ʉ ː ɹ i ə]
Occurrences:
7,208
Examples:
* pea:
[p ʰ i ː]
* gaea:
[d ʒ i ː ə]
* sui:
[s w i ː]
* eprix:
[i ː p ɹ i ː]
Occurrences:
2,738
Examples:
* whose:
[h ʉ ː z]
* pule:
[p ʲ ʉ ː ɫ]
* boot:
[b ʉ ː t]
* lula:
[ʎ ʉ l ə]
Occurrences:
5,399
Examples:
* toon:
[t ʰ ʉ ː n]
* roup:
[ɹ ʉ ː p]
* new:
[ɲ ʉ ː]
* spool:
[s p ʉ ː ɫ]
Occurrences:
55,252
Examples:
* min:
[m ʲ ɪ n]
* swing:
[s w ɪ ŋ]
* humid:
[ç ʉ ː m ʲ ɪ d]
* comic:
[k ʰ ɑ m ʲ ɪ k]
Occurrences:
2,805
Examples:
* jurat:
[d ʒ ʊ ɹ æ t]
* oopsy:
[ʊ p s i]
* nuqta:
[n ʊ k t ə]
* usual:
[j ʉ ː ʒ ʊ ɫ ̩]

Close-Mid
Occurrences:
14,915
Examples:
* inane:
[ɪ ɲ e j n]
* dkx:
[d ʲ i ː c ʰ e j ɛ k s]
* beta:
[b e j t ə]
* waco:
[w e j k o w]
Occurrences:
11,587
Examples:
* does:
[d o w z]
* juno:
[d ʒ ʉ n o w]
* demos:
[d ɛ m o w z]
* bayou:
[b a j o w]
Occurrences:
43,919
Examples:
* semla:
[s ɛ m l ə]
* agen:
[e j d ʒ ə n]
* aziz:
[ə z i ː z]
* karma:
[k ʰ ɑ ɹ m ə]
Occurrences:
11,800
Examples:
* rager:
[ɹ e j d ʒ ɚ]
* mater:
[m e j t ɚ]
* shiur:
[ʃ i ː ɚ]
* acter:
[æ k t ɚ]

Open-Mid
Occurrences:
21,670
Examples:
* cared:
[c ʰ ɛ ɹ d]
* faria:
[f ɛ ɹ i ə]
* ettin:
[ɛ t ʲ ɪ n]
* steck:
[s t ɛ k]
Occurrences:
4,068
Examples:
* serb:
[s ɝ b]
* apert:
[ə p ʰ ɝ t]
* foehn:
[f ɝ n]
* kirby:
[k ʰ ɝ b ʲ i]
Occurrences:
18,489
Examples:
* ascot:
[æ s k ɑ ː t]
* jan:
[d ʒ æ n]
* amply:
[æ m p ʲ ʎ i]
* gaz:
[ɡ æ z]
Occurrences:
8,383
Examples:
* huns:
[h ɐ n z]
* budge:
[b ɐ d ʒ]
* musk:
[m ɐ s k]
* hasta:
[h ɐ s t ə]

Open
Occurrences:
12,257
Examples:
* graz:
[ɡ ɹ ɑ ː t s]
* shock:
[ʃ ɑ k]
* fossa:
[f ɑ s ə]
* hark:
[h ɑ ɹ k]
Occurrences:
2,841
Examples:
* nawab:
[n ə w ɑ ː b]
* draff:
[d ɹ ɑ ː f]
* redox:
[ɹ ɪ d ɑ ː k s]
* lavas:
[l ɑ ː v ə s]
Occurrences:
6,255
Examples:
* cross:
[k ɹ ɒ s]
* soars:
[s ɒ ɹ z]
* arlon:
[ɑ ɹ l ɒ ː n]
* volvo:
[v ɒ ɫ v o w]
Occurrences:
1,776
Examples:
* falls:
[f ɒ ː ɫ z]
* staal:
[s t ɒ ː ɫ]
* vaut:
[v ɒ ː t]
* auks:
[ɒ ː k s]

Diphthongs#

  • aj

  • aw

  • ej

  • ow

  • ɔj

On this page
Show Source