Ukrainian MFA dictionary v2.0.0a#
@techreport{mfa_ukrainian_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Ukrainian MFA dictionary v2.0.0a},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Ukrainian/Ukrainian MFA dictionary v2_0_0a.html}},
year={2022},
month={May},
}
G2P models Acoustic models |
Installation#
Install from the MFA command line:
mfa model download dictionary ukrainian_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 Ukrainian transcripts.
This dictionary uses the MFA phone set for Ukrainian, and was used in training the Ukrainian 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: 14,213 Examples: * армій: [ɑ ɾ m ʲ i i] * брамі: [b ɾ ɑ m ʲ i] * хома: [x ɔ m ɑ] * рам'я: [ɾ ɑ m j ɐ] Occurrences: 1,513 Examples: * шумів: [ʃ ʊ m ʲ i u] * усміх: [ʊ s ʲ m ʲ i x] * домів: [d ̪ ɔ m ʲ i u] * обмін: [ɔ b m ʲ i n ̪] Occurrences: 12 Examples: |
Occurrences: 21,333 Examples: * вічну: [ʋ ʲ i t ʃ n ̪ ʊ] * ознак: [ɔ z ̪ n ̪ ɑ k] * пхну: [p x n ̪ ʊ] * кинув: [k e n ̪ u u] Occurrences: 447 Examples: * панна: [p ɑ n ̪ ː ɐ] * сонне: [s ̪ ɔ ɔ n ̪ ː e] * цінно: [t s ʲ i n ̪ ː ɔ] * цінне: [t s ʲ i n ̪ ː e] |
Occurrences: 6,126 Examples: * осені: [ɔ s ̪ ɛ ɲ i] * кинді: [k ɪ ɲ d ʲ i] * ніяк: [ɲ i j ɑ k] * гніві: [ɦ ɲ i ʋ i] Occurrences: 1,210 Examples: * вання: [ʋ ɑ ɲ ː ɐ] * рання: [ɾ ɑ ɲ ː ɐ] * цінні: [t s ʲ i ɲ ː i] * кінні: [c i ɲ ː i] |
|||||
Stop |
Occurrences: 16,046 Examples: * окріп: [o k ɾ ʲ i p] * посуд: [p ɔ s ̪ ʊ d ̪] * печі: [p e t ʃ ʲ i] * прошу: [p ɾ ɔ ʃ ʊ] Occurrences: 7,165 Examples: * граби: [ɦ ɾ ɐ b ɪ] * любе: [ʎ ʊ b ɛ] * брови: [b ɾ ɔ ʋ ɪ] * багно: [b ɐ ɦ n ̪ ɔ] Occurrences: 3 Examples: |
Occurrences: 18,648 Examples: * стиль: [s ̪ t ̪ e ʎ] * плоту: [p l ɔ t ̪ ʊ] * канат: [k ɐ n ̪ ɑ t ̪] * діти: [d ʲ i t ̪ ɪ] Occurrences: 20 Examples: * гетто: [ɦ ɛ t ̪ ː ɔ] Occurrences: 11,543 Examples: * дзюдо: [d z ʲ ʊ d ̪ ɔ] * заїду: [z ̪ ɐ j i d ̪ ʊ] * анди: [ɐ n ̪ d ̪ ɪ] * льоду: [ʎ ɔ d ̪ ʊ] Occurrences: 122 Examples: * будда: [b u d ̪ ː ɐ] * міддю: [m ʲ i d ̪ ː ʊ] * оддам: [ɔ d ̪ ː ɐ m] * оддай: [ɔ d ̪ ː ɑ i] |
Occurrences: 1,162 Examples: * кібця: [c i b t s ʲ ɐ] * кіоск: [c i ɔ s ̪ k] * кім: [c i m] * кіз: [c i z ̪] Occurrences: 1 Examples: |
Occurrences: 17,954 Examples: * боках: [b ɔ k ɐ x] * муки: [m u k ɪ] * котив: [k ɔ t ̪ ɪ u] * ящику: [j ɑ ʃ t ʃ e k ʊ] Occurrences: 8 Examples: * мекку: [m ɛ k ː ʊ] * мекка: [m ɛ k ː ɐ] Occurrences: 134 Examples: * ґпу: [ɡ p ʊ] * аґрус: [ɑ ɡ ɾ ʊ s ̪] * ґазда: [ɡ ɑ z ̪ d ̪ ɑ] * манґи: [m ɐ n ̪ ɡ ɪ] Occurrences: 1 Examples: * меґґі: [m e ɡ ː i] |
||||
Affricate |
Occurrences: 952 Examples: * шприц: [ʃ p ɾ e t ̪ s ̪] * лице: [l e t ̪ s ̪ ɛ] * цирк: [t ̪ s ̪ ɪ ɾ k] * царям: [t ̪ s ̪ ɑ ɾ ɐ m] Occurrences: 4 Examples: * цска: [t ̪ s ̪ ː k ɐ] Occurrences: 418 Examples: * будз: [b u d ̪ z ̪] * гудзя: [ɦ u d ̪ z ̪ ɐ] * дзиґа: [d ̪ z ̪ ɪ ɡ ɐ] * дзень: [d ̪ z ̪ ɛ ɲ] Occurrences: 1 Examples: |
Occurrences: 7,635 Examples: * схочу: [s ̪ x ɔ t ʃ ʊ] * вечір: [ʋ ɛ t ʃ ʲ i ɾ] * чесав: [t ʃ e s ̪ ɑ u] * чорна: [t ʃ ɔ ɾ n ̪ ɐ] Occurrences: 97 Examples: * матч: [m ɐ t ʃ ː] * матчу: [m ɐ t ʃ ː ʊ] * одчай: [ɔ t ʃ ː ɐ i] * одчув: [ɔ t ʃ ː u u] Occurrences: 579 Examples: * адже: [ɐ d ʒ e] * раджу: [ɾ ɐ d ʒ ʊ] * джаз: [d ʒ ɑ z ̪] * джолі: [d ʒ ɔ ʎ i] |
||||||
Sibilant |
Occurrences: 13,557 Examples: * осока: [ɔ s ̪ ɔ k ɐ] * совок: [s ̪ ɔ ʋ ɔ k] * покіс: [p o c i s ̪] * синка: [s ̪ ɪ n ̪ k ɐ] Occurrences: 25 Examples: * ссати: [s ̪ ː ɑ t ̪ ɪ] * улісс: [ʊ ʎ i s ̪ ː] * ссав: [s ̪ ː ɐ u] * цска: [t ̪ s ̪ ː k ɐ] Occurrences: 9,752 Examples: * знав: [z ̪ n ̪ ɑ u] * назв: [ɛ n ̪ ɑ z ̪ ɛ ʋ ɛ] * муз: [m u z ̪] * щезли: [ʃ t ʃ ɛ z ̪ l ɪ] Occurrences: 25 Examples: * ззаду: [z ̪ ː ɑ d ̪ ʊ] * ззаді: [z ̪ ː ɑ d ʲ i] |
Occurrences: 10,590 Examples: * касян: [k ɐ s ʲ ɑ n ̪] * лайся: [l ɐ i s ʲ ɐ] * сядь: [s ʲ ɑ d ʲ] * хижці: [x e ʒ t s ʲ i] Occurrences: 109 Examples: * хатці: [x ɑ ɔ t s ʲ ː i] * кітці: [c i ɔ t s ʲ ː i] * дочці: [d ̪ ɔ t s ʲ ː i] * людці: [ʎ ʊ t s ʲ ː i] Occurrences: 1,553 Examples: * взято: [ʋ z ʲ ɑ t ̪ ɔ] * дузі: [d ̪ u z ʲ i] * черзі: [t ʃ ɛ ɾ z ʲ i] * стезі: [s ̪ t ̪ e z ʲ i] Occurrences: 8 Examples: |
Occurrences: 5,968 Examples: * щоку: [ʃ t ʃ o k u] * шори: [ʃ ɔ ɾ ɪ] * чував: [t ʃ ʊ ʋ ɑ u] * чарку: [t ʃ ɑ ɾ k ʊ] Occurrences: 151 Examples: * марші: [m ɑ ɾ ʃ ʲ i] * почім: [p o t ʃ ʲ i m] * прищі: [p ɾ e ʃ t ʃ ʲ i] * дужчі: [d ̪ u ʒ t ʃ ʲ i] Occurrences: 3 Examples: * ніччю: [ɲ i t ʃ ʲ ː ʊ] * матчі: [m ɐ t ʃ ʲ ː i] * рітчі: [ɾ ʲ i t ʃ ʲ ː i] * віччю: [ʋ ʲ i t ʃ ʲ ː ʊ] Occurrences: 3,304 Examples: * гожа: [ɦ ɔ ʒ ɐ] * жахом: [ʒ ɑ x ɔ m] * живеш: [ʒ e ʋ ɛ ʃ] * жене: [ʒ e n ̪ ɛ] Occurrences: 68 Examples: * жінок: [ʒ ʲ i n ̪ ɔ k] * жінку: [ʒ ʲ i n ̪ k ʊ] * жінко: [ʒ ʲ i n ̪ k ɔ] * жінці: [ʒ ʲ i ɲ t s ʲ i] Occurrences: 23 Examples: |
|||||
Fricative |
Occurrences: 839 Examples: * цифри: [t ̪ s ̪ ɪ f ɾ ɪ] * моффа: [m ɔ f f ɐ] * нафті: [n ̪ ɑ f t ̪ i] * німф: [ɲ i m f] Occurrences: 341 Examples: * софія: [s ̪ o f ʲ i j ɐ] * рифів: [ɾ e f ʲ i u] * фіна: [f ʲ i n ̪ ɐ] * фірм: [f ʲ i ɾ m] |
Occurrences: 233 Examples: * хітом: [ç i t ̪ ɔ m] * хіть: [ç i t ʲ] * хіба: [ç i b ɑ] * вхід: [ʋ ç i d ̪] Occurrences: 518 Examples: * гіди: [ʝ i d ̪ ɪ] * агій: [ɐ ʝ i i] * гірка: [ʝ i ɾ k ɐ] * легіт: [l e ʝ i t ̪] |
Occurrences: 8,318 Examples: * догм: [d ̪ ɔ ɦ m] * горді: [ɦ ɔ ɾ d ʲ i] * гутан: [ɦ ʊ t ̪ ɑ n ̪] * цього: [t s ʲ ɔ ɦ ɔ] Occurrences: 5 Examples: * реггі: [ɾ e ɦ ː i] |
|||||
Approximant |
Occurrences: 16,744 Examples: * вузи: [ʋ u z ̪ ɪ] * вої: [ʋ ɔ j i] * всею: [ʋ s ̪ e j ʊ] * внесу: [ʋ n ̪ ɛ s ̪ ʊ] Occurrences: 3,610 Examples: * звіту: [z ʲ ʋ ʲ i t ̪ ʊ] * віза: [ʋ ʲ i z ̪ ɐ] * повіє: [p ɔ ʋ ʲ i j e] * свіі: [s ʲ ʋ ʲ i i] Occurrences: 21 Examples: * ввів: [ʋ ʲ ː i u] * вві: [ʋ ʲ ː i] Occurrences: 47 Examples: * ввело: [ʋ ː e l ɔ] * ввдно: [ʋ ː d ̪ n ̪ ɔ] * вволю: [ʋ ː ɔ ʎ ʊ] * ввели: [ʋ ː e l ɪ] |
Occurrences: 9,836 Examples: * воєві: [ʋ ɔ j e ʋ ʲ i] * ситої: [s ̪ e t ̪ ɔ j i] * темою: [t ̪ e m ɔ j ʊ] * устає: [ʊ s ̪ t ̪ ɑ j e] |
||||||
Tap |
Occurrences: 24,195 Examples: * радує: [ɾ ɑ d ̪ ʊ j e] * керчі: [k ɛ ɾ t ʃ ʲ i] * царка: [t ̪ s ̪ ɑ ɾ k ɐ] * юрасю: [j ʊ ɾ ɐ s ʲ u] Occurrences: 2,826 Examples: * пряла: [p ɾ ʲ ɑ l ɐ] * котрі: [k o t ʲ ɾ ʲ i] * рідше: [ɾ ʲ i d ̪ ʃ e] * нарік: [n ̪ ɑ ɾ ʲ i k] Occurrences: 6 Examples: * гаррі: [ɦ ɐ ɾ ʲ ː i] Occurrences: 14 Examples: * ферро: [f ɛ ɾ ː ɔ] * гурра: [ɦ ʊ ɾ ː ɐ] |
|||||||
Lateral |
Occurrences: 16,187 Examples: * пекли: [p ɛ k l ɪ] * клеєм: [k l ɛ ɛ m] * олово: [ɔ l ɔ ʋ ɔ] * ланах: [l ɐ n ̪ ɐ x] Occurrences: 15 Examples: * алло: [ɐ l ː ɔ] * алмаз: [ɐ l ː m ɑ z ̪] * аллах: [ɐ l ː ɑ x] * аллен: [ɐ l ː ɛ n ̪] |
Occurrences: 6,047 Examples: * горлі: [ɦ ɔ ɾ ʎ i] * рубля: [ɾ ʊ b ʎ ɐ] * любці: [ʎ ʊ b t s ʲ i] * людях: [ʎ u d ̪ ɐ x] Occurrences: 93 Examples: * рілля: [ɾ ʲ i ʎ ː ɐ] * заллє: [z ̪ ɐ ʎ ː ɛ] * гіллі: [ʝ i ʎ ː i] * зілля: [z ʲ i ʎ ː ɐ] |
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: 32,400 Examples: * вірно: [ʋ ʲ i ɾ n ̪ ɔ] * голій: [ɦ ɔ ʎ i i] * їду: [j i d ̪ ʊ] * ідей: [i d ̪ ɛ i] |
Occurrences: 14,200 Examples: * пудру: [p u d ̪ ɾ ʊ] * сікур: [s ʲ i k u ɾ] * авеню: [ɐ ʋ e ɲ u] * усик: [u s ̪ e k] |
|||
Occurrences: 24,765 Examples: * тьми: [t ʲ m ɪ] * горби: [ɦ ɔ ɾ b ɪ] * шахи: [ʃ ɑ x ɪ] * утіхи: [u t ʲ i x ɪ] |
Occurrences: 17,420 Examples: * курку: [k ʊ ɾ k ʊ] * шапку: [ʃ ɑ p k ʊ] * посуд: [p ɔ s ̪ ʊ d ̪] * гуку: [ɦ ʊ k ʊ] |
||||
Close-Mid |
Occurrences: 31,256 Examples: * влізе: [ʋ ʎ i z ̪ e] * текти: [t ̪ e k t ̪ ɪ] * серед: [s ̪ ɛ ɾ e d ̪] * кисне: [k ɪ s ̪ n ̪ e] |
Occurrences: 2,584 Examples: * росія: [ɾ o s ʲ i j ɐ] * боку: [b o k u] * копій: [k o p ʲ i i] * болі: [b o ʎ i] |
|||
Open-Mid |
Occurrences: 7,650 Examples: * ввесь: [ʋ ː ɛ s ʲ] * імен: [i m ɛ n ̪] * сесю: [s ̪ ɛ s ʲ u] * менко: [m ɛ n ̪ k ɔ] |
Occurrences: 40,873 Examples: * любої: [ʎ ʊ b ɔ j i] * торги: [t ̪ ɔ ɾ ɦ ɪ] * двору: [d ̪ ʋ ɔ ɾ ʊ] * него: [n ̪ e ɦ ɔ] |
|||
Occurrences: 33,656 Examples: * змила: [z ̪ m ɪ l ɐ] * шмата: [ʃ m ɑ t ̪ ɐ] * заява: [z ̪ ɐ j ɑ ʋ ɐ] * срака: [s ̪ ɾ ɑ k ɐ] |
|||||
Open |
Occurrences: 23,214 Examples: * косак: [k ɔ s ̪ ɑ k] * кусай: [k ʊ s ̪ ɑ i] * женя: [ʒ e ɲ ɑ] * дяче: [d ʲ ɑ t ʃ e] |