English (US) ARPA dictionary v2.0.0#
@article{gorman2011prosodylab,
author={Gorman, Kyle and Howell, Jonathan and Wagner, Michael},
title={Prosodylab-aligner: A tool for forced alignment of laboratory speech},
journal={Canadian Acoustics},
volume={39},
number={3},
pages={192--193},
year={2011}
}
G2P models Acoustic models |
Installation#
Install from the MFA command line:
mfa model download dictionary english_us_arpa
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 ARPA phone set for English, and was used in training the English ARPA 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: 44,595 Examples: * visum: [V IH1 ZH AH0 M] * mako: [M AA1 K OW0] * myrt: [M ER1 T] * alm: [AA1 L M] |
Occurrences: 94,848 Examples: * unlit: [AH0 N L IH2 T] * gans: [G AE1 N Z] * lunga: [L AH1 N G AH0] * indo: [IH1 N D OW0] |
Occurrences: 15,230 Examples: * sangs: [S AE1 NG Z] * sings: [S IH1 NG Z] * tanga: [T AA1 NG G AH0] * stank: [S T AE1 NG K] |
|||||
Stop |
Occurrences: 32,322 Examples: * p'hra: [P IY1 R AH0] * pabo: [P AA1 B OW0] * paht: [P AA1 T] * soupe: [S UW1 P] Occurrences: 31,562 Examples: * haub: [HH AO1 B] * pabo: [P AA1 B OW0] * bhuja: [B UW1 Y AH0] * babin: [B AE1 B IH0 N] |
Occurrences: 75,542 Examples: * unlit: [AH0 N L IH2 T] * corta: [K AO1 R T AH0] * myrt: [M ER1 T] * sart: [S AA0 R T] Occurrences: 52,334 Examples: * herod: [HH EH1 R AH0 D] * indo: [IH1 N D OW0] * dour: [D AW1 ER0] * dewy: [D UW1 IY0] |
Occurrences: 58,375 Examples: * coru: [K AO1 R UW0] * corta: [K AO1 R T AH0] * mako: [M AA1 K OW0] * saxo: [S AE1 K S OW0] Occurrences: 20,766 Examples: * gans: [G AE1 N Z] * lunga: [L AH1 N G AH0] * gyu: [G Y UW2] * agnew: [AE1 G N Y UW0] |
|||||
Affricate |
Occurrences: 8,420 Examples: * check: [CH EH1 K] * chaf: [CH AE1 F] * tchah: [CH] * vich: [V IH0 CH] Occurrences: 10,111 Examples: * jarl: [JH AA1 R L] * monge: [M AA1 N JH] * rigi: [R IH1 JH IY0] * job's: [JH AA1 B Z] |
|||||||
Sibilant |
Occurrences: 78,674 Examples: * sangs: [S AE1 NG Z] * saxo: [S AE1 K S OW0] * sart: [S AA0 R T] * areus: [EH1 R IY0 AH0 S] Occurrences: 53,038 Examples: * sangs: [S AE1 NG Z] * gans: [G AE1 N Z] * norms: [N AO1 R M Z] * lawns: [L AO1 N Z] |
Occurrences: 12,288 Examples: * tisza: [T IH1 SH AH0] * shark: [SH AA1 R K] * imshi: [IH1 M SH IY0] * fosh: [F AA1 SH] Occurrences: 796 Examples: * visum: [V IH1 ZH AH0 M] * touge: [T UW1 ZH] * zhoo: [ZH UW1] * taj: [T AA1 ZH] |
||||||
Fricative |
Occurrences: 20,957 Examples: * fout: [F AW1 T] * frey: [F R EY1] * chaf: [CH AE1 F] * flint: [F L IH1 N T] Occurrences: 17,526 Examples: * visum: [V IH1 ZH AH0 M] * vitta: [V IY1 T AH0] * giv: [G IH1 V] * visby: [V IH1 S B IY0] |
Occurrences: 6,592 Examples: * seeth: [S IY1 TH] * nith: [N IH2 TH] * thud: [TH AH1 D] * oneth: [W AH1 N EH1 TH] Occurrences: 1,311 Examples: * this: [DH IH1 S] * lathe: [L EY1 DH] * than: [DH AE1 N] * otha: [AH0 DH AA1] |
Occurrences: 13,675 Examples: * haub: [HH AO1 B] * herod: [HH EH1 R AH0 D] * aloha: [AH0 L OW1 HH AA0] * helve: [HH EH1 L V] |
|||||
Approximant |
Occurrences: 14,822 Examples: * wixy: [W IH1 K S IY0] * aiwa: [AY1 W AH0] * swat: [S W AA1 T] * dwell: [D W EH1 L] |
Occurrences: 71,342 Examples: * coru: [K AO1 R UW0] * corta: [K AO1 R T AH0] * p'hra: [P IY1 R AH0] * sart: [S AA0 R T] |
Occurrences: 8,159 Examples: * bhuja: [B UW1 Y AH0] * gyu: [G Y UW2] * agnew: [AE1 G N Y UW0] * humes: [HH Y UW1 M Z] |
|||||
Lateral |
Occurrences: 76,788 Examples: * unlit: [AH0 N L IH2 T] * alm: [AA1 L M] * malus: [M AH0 L AH0 S] * lunga: [L AH1 N G AH0] |
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: 55,169 Examples: * kee: [K IY1] * meder: [M IY1 D ER0] * romae: [R OW0 M IY2] * kwee: [K W IY2] |
Occurrences: 16,256 Examples: * bhuja: [B UW1 Y AH0] * coru: [K AO1 R UW0] * pouce: [P UW2 S] * dewy: [D UW1 IY0] |
|||
Occurrences: 79,698 Examples: * critz: [K R IH1 T S] * repel: [R IH0 P EH1 L] * istar: [IH2 S T AA0 R] * stip: [S T IH2 P] |
Occurrences: 4,024 Examples: * hould: [HH UH1 D] * suraj: [S UH0 R AA1 ZH] * cures: [K Y UH1 R Z] * toots: [T UH1 T S] |
||||
Close-Mid |
Occurrences: 21,225 Examples: * aoste: [EY1 OW1 S T] * mamey: [M EY1 M IY0] * nacre: [N EY2 K ER0] * whei: [HH W EY2] |
Occurrences: 25,875 Examples: * mako: [M AA1 K OW0] * leggo: [L EH1 G OW2] * saxo: [S AE1 K S OW0] * aoste: [EY1 OW1 S T] |
|||
Occurrences: 116,953 Examples: * p'hra: [P IY1 R AH0] * suf: [S AH2 F] * k'ung: [K EY1 SH AH2 NG] * suld: [S AH1 L D] |
|||||
Open-Mid |
Occurrences: 39,703 Examples: * bicep: [B AY2 S EH0 P] * etext: [EH2 T EH1 K S T] * ngalo: [EH0 NG G AH0 L OW2] * abdel: [AE1 B D EH2 L] |
||||
Occurrences: 34,206 Examples: * imlac: [IH0 M L AE2 K] * gans: [G AE1 N Z] * akhab: [AE1 K AE2 B] * babin: [B AE1 B IH0 N] |
Occurrences: 40,624 Examples: * carus: [K ER2 AH0 S] * boker: [B OW1 K ER0] * nacre: [N EY2 K ER0] * myrt: [M ER1 T] |
||||
Open |
Occurrences: 37,763 Examples: * narth: [N AA2 R TH] * otkar: [AA1 T K AA2 R] * pabo: [P AA1 B OW0] * claro: [K L AA1 R OW0] Occurrences: 18,014 Examples: * corta: [K AO1 R T AH0] * autou: [AO2 T UW1] * gorp: [G AO0 R P] * coru: [K AO1 R UW0] |
Diphthongs#
AW
AY
OY
Stress#
0
1
2