Hausa MFA dictionary v2.0.0#
@techreport{mfa_hausa_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Hausa MFA dictionary v2.0.0},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Hausa/Hausa MFA dictionary v2_0_0.html}},
year={2022},
month={Mar},
}
G2P models Acoustic models |
Installation#
Install from the MFA command line:
mfa model download dictionary hausa_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 Hausa transcripts.
This dictionary uses the MFA phone set for Hausa, and was used in training the Hausa 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 |
Alveolar |
Alveopalatal |
Retroflex |
Palatal |
Velar |
Glottal |
---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 1,516 Examples: * komo: [kʷ oː˥ m oː˥] * meg: [m eː˥ ɡ] * miƙe: [m ɪ˥ cʼ eː˩] * gamu: [ɡ aː˥ m ʊ˩] |
Occurrences: 1,703 Examples: * anzu: [ʔ a˩ n z ɪ˩] * hana: [h aː˥ n aː˩] * ɗina: [ɗ ɪ˥ n aː˩] * anaca: [ʔ a˩ n ə˩ tʃ ə˩] |
Occurrences: 106 Examples: * tsiya: [sʼ ɪ˥ ɲ j aː˥] * yanyi: [j ɛ˥ ɲ j iː˥] * hange: [h ɛ˥ ɲ ɟ eː˥] * sanye: [s ɛ˥ ɲ j eː˥] |
Occurrences: 1,559 Examples: * jin: [dʒ ɪ˥ ŋ] * sojan: [s oː˥ dʒ a˥ ŋ] * ilmin: [ʔ ɪ˥ l m ɪ˥ ŋ] * nufin: [n ɪ˥ ɸ ɪ˩ ŋ] |
|||
Stop Plain |
Occurrences: 83 Examples: * pan: [p a˥˦ ŋ] * trump: [t ɽ ɪ˥ m p] * pine: [p ɪ˩ n e˥] * bappa: [b a˥ p p ə˩] Occurrences: 1,026 Examples: * bas: [b a˥˦ s] * binta: [b ɪ˥ n t ə˩] * bude: [b ʊ˩ d e˩] * abin: [ʔ a˩ b ɪ˩ ŋ] |
Occurrences: 1,276 Examples: * fita: [ɸ ɪ˩ t aː˥] * binta: [b ɪ˥ n t ə˩] * atiku: [ʔ a˩ t ɪ˩ kʷ ʊ˩] * taha: [t ə˥ h aː˩] Occurrences: 991 Examples: * gada: [ɡ ə˥ d aː˩] * ƙundu: [kʷʼ ʊ˩ n d uː˥] * hadu: [h a˩ d ɪ˩] * bude: [b ʊ˩ d e˩] |
Occurrences: 469 Examples: * kyale: [c ɛ˥ l eː˩] * dake: [d ə˩ c eː˥] * kek: [c eː˥ k] * ken: [c eː˥ ŋ] Occurrences: 159 Examples: * rage: [r ə˥ ɟ eː˥] * bige: [b ɪ˥ ɟ e˥] * begen: [b a˥ ɟ eː˥ ŋ] * kogon: [kʷ oː˩ ɟ ɔ˩ ŋ] |
Occurrences: 901 Examples: * kauce: [k aw˥ tʃ eː˩] * mukai: [m uː˥ k aj˥] * kek: [c eː˥ k] * nunka: [n ɪ˩ ŋ k aː˥] Occurrences: 359 Examples: * gada: [ɡ ə˥ d aː˩] * meg: [m eː˥ ɡ] * gaɗa: [ɡ ə˥ ɗ aː˩] * loga: [l oː˥ ɡ aː˩] |
Occurrences: 910 Examples: * anzu: [ʔ a˩ n z ɪ˩] * anaca: [ʔ a˩ n ə˩ tʃ ə˩] * ilmin: [ʔ ɪ˥ l m ɪ˥ ŋ] * atiku: [ʔ a˩ t ɪ˩ kʷ ʊ˩] |
||
Implosive |
Occurrences: 67 Examples: * haɓa: [h a˥ ɓ aː˩] * zaɓin: [z aː˥ ɓ ɪ˩ ŋ] * goɓe: [ɡʷ oː˩ ɓ e˩] * karɓi: [k a˥ r ɓ ɪ˩] |
Occurrences: 145 Examples: * ɗina: [ɗ ɪ˥ n aː˩] * gaɗa: [ɡ ə˥ ɗ aː˩] * faɗi: [ɸ aː˥ ɗ iː˥] * ɗa'a: [ɗ aː˥ ʔ aː˩] |
|||||
Affricate |
Occurrences: 475 Examples: * kauce: [k aw˥ tʃ eː˩] * anaca: [ʔ a˩ n ə˩ tʃ ə˩] * soci: [s oː˥ tʃ ɪ˩] * cao: [tʃ o˥] Occurrences: 488 Examples: * jaje: [dʒ aː˥ dʒ e˩] * jin: [dʒ ɪ˥ ŋ] * sojan: [s oː˥ dʒ a˥ ŋ] * jawur: [dʒ aː˩ w ʊ˥ r] |
||||||
Sibilant Plain |
Occurrences: 1,288 Examples: * sauƙi: [s aw˩ cʼ iː˩] * bas: [b a˥˦ s] * sojan: [s oː˥ dʒ a˥ ŋ] * sura: [s uː˥ r aː˩] Occurrences: 328 Examples: * anzu: [ʔ a˩ n z ɪ˩] * gwaza: [ɡʷ aː˥ z aː˥] * zaɓin: [z aː˥ ɓ ɪ˩ ŋ] * zemin: [z eː˥ m ɪ˥ ŋ] |
Occurrences: 399 Examples: * sashi: [s ə˩ ʃ iː˥] * sheƙa: [ʃ eː˩ kʼ aː˩] * sha: [ʃ ə˥] * toh: [t oː˥ ʃ] |
|||||
Ejective |
Occurrences: 195 Examples: * yatsu: [j aː˥ sʼ uː˥] * tsiya: [sʼ ɪ˥ ɲ j aː˥] * tsit: [sʼ ɪ˥ t] * tsaya: [sʼ ɛ˥ j aː˩] |
||||||
Fricative |
Occurrences: 630 Examples: * fita: [ɸ ɪ˩ t aː˥] * faɗi: [ɸ aː˥ ɗ iː˥] * nufin: [n ɪ˥ ɸ ɪ˩ ŋ] * feng: [ɸ ɛ˥ ŋ] |
Occurrences: 655 Examples: * hana: [h aː˥ n aː˩] * hali: [h aː˥ l iː˥] * hadu: [h a˩ d ɪ˩] * haɓa: [h a˥ ɓ aː˩] |
|||||
Approximant |
Occurrences: 831 Examples: * jawur: [dʒ aː˩ w ʊ˥ r] * jawad: [dʒ ə˩ w aː˥ d] * wayo: [w aː˥ j oː˥˦] * bawa: [b aː˥ w aː˩] |
Occurrences: 1,044 Examples: * yuli: [j uː˥ l ɪ˩] * rayar: [ɽ aː˩ j ɛ˥ r] * yana: [j aː˥ n aː˩] * koya: [kʷ oː˩ j aː˩] |
|||||
Tap |
Occurrences: 742 Examples: * roger: [ɽ oː˥ ɡʷ eː˥ r] * kurma: [kʷ ʊ˥ ɽ m aː˥] * ƙarar: [kʼ ə˩ ɽ r] * rayar: [ɽ aː˩ j ɛ˥ r] |
||||||
Trill |
Occurrences: 1,679 Examples: * roger: [ɽ oː˥ ɡʷ eː˥ r] * ƙafar: [kʼ ə˥ ɸ a˩ r] * jawur: [dʒ aː˩ w ʊ˥ r] * sura: [s uː˥ r aː˩] |
||||||
Lateral |
Occurrences: 1,135 Examples: * yuli: [j uː˥ l ɪ˩] * loga: [l oː˥ ɡ aː˩] * hali: [h aː˥ l iː˥] * lao: [l a˥ o˥] |
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: 1,257 Examples: * buri: [b ʊ˩ r iː˩] * kari: [k ə˩ r iː˥] * tasi: [t aː˩ s iː˥˦] * sauti: [s aw˥ t iː˩] |
Occurrences: 520 Examples: * huhu: [h uː˥ h uː˩] * you: [j oː˥˦ uː˩] * hula: [h uː˩ l aː˥] * ƙundu: [kʷʼ ʊ˩ n d uː˥] |
|||
Occurrences: 3,012 Examples: * hadu: [h a˩ d ɪ˩] * ɗina: [ɗ ɪ˥ n aː˩] * yuli: [j uː˥ l ɪ˩] * binta: [b ɪ˥ n t ə˩] |
Occurrences: 1,010 Examples: * jawur: [dʒ aː˩ w ʊ˥ r] * juna: [dʒ ʊ˩ n aː˥] * ƙundu: [kʷʼ ʊ˩ n d uː˥] * gamu: [ɡ aː˥ m ʊ˩] |
||||
Close-Mid |
Occurrences: 183 Examples: * bude: [b ʊ˩ d e˩] * bige: [b ɪ˥ ɟ e˥] * sef: [s e˥˦ ɸ] * guje: [ɡʷ ʊ˩ dʒ e˩] Occurrences: 869 Examples: * miƙe: [m ɪ˥ cʼ eː˩] * dake: [d ə˩ c eː˥] * roger: [ɽ oː˥ ɡʷ eː˥ r] * meg: [m eː˥ ɡ] |
Occurrences: 88 Examples: * fasfo: [ɸ a˩ s ɸ o˥] * faɗo: [ɸ aː˥ ɗ o˩] * diego: [d ɪ˥ eː˥ ɡʷ o˩] * ceto: [tʃ eː˩ t o˥] Occurrences: 950 Examples: * iyo: [ʔ ɪ˩ j oː˥˦] * sojan: [s oː˥ dʒ a˥ ŋ] * koya: [kʷ oː˩ j aː˩] * you: [j oː˥˦ uː˩] |
|||
Occurrences: 2,453 Examples: * gaɗa: [ɡ ə˥ ɗ aː˩] * taha: [t ə˥ h aː˩] * gada: [ɡ ə˥ d aː˩] * binta: [b ɪ˥ n t ə˩] |
|||||
Open-Mid |
Occurrences: 454 Examples: * kyale: [c ɛ˥ l eː˩] * biyan: [b ɪ˥ j ɛ˩ ŋ] * rayar: [ɽ aː˩ j ɛ˥ r] * yakan: [j ɛ˥ k a˩ ŋ] |
Occurrences: 55 Examples: * sol: [s ɔ˥ l] * dokar: [d ɔ˥ k a˩ r] * manjo: [m a˥ n dʒ ɔ˩] * gabon: [ɡ ə˥ b ɔ˩ ŋ] |
|||
Open |
Occurrences: 2,650 Examples: * ƙafar: [kʼ ə˥ ɸ a˩ r] * anzu: [ʔ a˩ n z ɪ˩] * anaca: [ʔ a˩ n ə˩ tʃ ə˩] * haɓa: [h a˥ ɓ aː˩] Occurrences: 3,783 Examples: * abada: [ʔ a˩ b ə˩ d aː˥˦] * iraƙi: [ʔ ɪ˩ r aː˥˦ cʼ ɪ˩] * ɗina: [ɗ ɪ˥ n aː˩] * asuba: [ʔ a˩ s ɪ˩ b aː˥˦] |
Diphthongs#
aj
aw
Tones#
˥
˥˦
˩