Portuguese MFA dictionary v2.0.0a#
@techreport{mfa_portuguese_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Portuguese MFA dictionary v2.0.0a},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Portuguese/Portuguese MFA dictionary v2_0_0a.html}},
year={2022},
month={May},
}
Acoustic models |
Installation#
Install from the MFA command line:
mfa model download dictionary portuguese_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 Portuguese transcripts.
This dictionary uses the MFA phone set for Portuguese, and was used in training the Portuguese 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 |
Uvular |
---|---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 25,995 Examples: * mores: [m o ɾ ɨ ʃ] * mília: [m i l j ɐ] * gume: [ɡ u m e] * mana: [m ɐ n ɐ] |
Occurrences: 16,918 Examples: * qüen: [k w ɛ n] * nevão: [n ɨ v ɐ ̃ w ̃] * nead: [n j ɐ d] * nop: [n ɔ p] |
Occurrences: 4,337 Examples: * canil: [k ɐ ɲ i w] * toni: [t o ɲ i] * wayne: [w e j ɲ i] * cahos: [k ɐ ɲ u ʃ] |
|||||
Stop |
Occurrences: 25,647 Examples: * keep: [k ɨ p] * pardo: [p a x d u] * pem: [p ɐ ̃ j ̃] * tupis: [t u p i ʃ] Occurrences: 10,245 Examples: * bunda: [b u ̃ d ɐ] * bodum: [b o ð u ̃] * bilu: [b i l u] * bolo: [b o l u] |
Occurrences: 46,035 Examples: * peste: [p ɛ s t ʃ i] * ivete: [i v ɛ t ʃ i] * altar: [a w t a x] * tíbia: [t ʃ i b i ɐ] Occurrences: 24,794 Examples: * seda: [s e d a] * aguda: [a ɡ u d ɐ] * nodoa: [n ɔ d o ɐ] * decai: [d ɨ k a j] |
Occurrences: 2,458 Examples: * cnpq: [s e e ɲ i p e c e] * caric: [k ɐ ɾ i c i] * kiev: [c i ɛ v] * quito: [c i t u] Occurrences: 664 Examples: * gate: [ɟ e j t ʃ] * ong: [o ̃ ɟ i] * guias: [ɟ i ɐ s] * negue: [n e ɟ i] |
Occurrences: 31,786 Examples: * coado: [k w ɐ ð u] * cujo: [k u ʒ u] * carré: [k ɐ ʁ ɛ] * curou: [k u ɾ o w] Occurrences: 7,028 Examples: * tango: [t ɐ ̃ ɡ u] * yoga: [j ɔ ɡ a] * gogol: [ɡ o ɡ ɔ l] * gomma: [ɡ o m u m ɐ] |
||||
Affricate |
Occurrences: 4,723 Examples: * kut: [k u t ʃ] * latet: [l ɐ t ɛ t ʃ i] * brat: [b ɾ a t ʃ i] * tweet: [t w i t ʃ i] Occurrences: 2,952 Examples: * ardia: [a x d ʒ i ɐ] * pediu: [p i d ʒ i w] * pedia: [p e d ʒ i ɐ] * led: [l ɛ d ʒ] |
|||||||
Sibilant |
Occurrences: 44,613 Examples: * ossea: [ɔ s j ɐ] * serie: [s ɛ ɾ j ɨ] * sub: [s u b i] * ácida: [a s i d ɐ] Occurrences: 12,660 Examples: * resis: [ʁ ɨ z i s] * elisá: [i l i z a] * alisa: [ɐ l i z ɐ] * asul: [ɐ z u l] |
Occurrences: 37,775 Examples: * idas: [i ð ɐ ʃ] * ais: [a j ʃ] * leles: [l e l ɨ ʃ] * elmos: [ɛ l m u ʃ] Occurrences: 8,369 Examples: * onde: [o ̃ d ʒ] * aja: [a ʒ ɐ] * rége: [ʁ ɛ ʒ ɨ] * canja: [k ɐ ̃ ʒ a] |
||||||
Fricative |
Occurrences: 13,090 Examples: * frapê: [f ɾ a p e] * frade: [f ɾ a ð ɨ] * fás: [f a ʃ] * féras: [f ɛ ɾ ɐ ʃ] Occurrences: 17,629 Examples: * povna: [p ɔ v n ɐ] * otava: [o t a v ɐ] * varra: [v a ʁ ɐ] * sirva: [s i ɾ v ɐ] |
Occurrences: 16,620 Examples: * caído: [k ɐ i ð u] * nódoa: [n ɔ ð o ɐ] * vozde: [v ɔ z ð ɨ] * sêdas: [s e ð ɐ ʃ] |
Occurrences: 9,875 Examples: * darei: [d a ʁ e j] * rale: [ʁ ɐ l ɨ] * rucha: [ʁ u ʃ ɐ] * rolou: [ʁ o l o w] |
|||||
Approximant |
Occurrences: 12,894 Examples: * rol: [x ɔ w] * cocal: [k o k a w] * golpe: [ɡ ɔ w p i] * leal: [l e a w] Occurrences: 8,524 Examples: * spam: [s p ɐ ̃ w ̃] * quão: [k w ɐ ̃ w ̃] * irão: [i ɾ ɐ ̃ w ̃] * facão: [f a k ɐ ̃ w ̃] |
Occurrences: 21,706 Examples: * impôs: [i ̃ p o j s] * jânio: [ʒ ɐ n j u] * hater: [e j t e ɾ] * talia: [t ɐ l j ɐ] Occurrences: 4,971 Examples: * além: [a l e ̃ j ̃] * uem: [ɐ ̃ j ̃] * modem: [m o d e ̃ j ̃] * latem: [l a t e ̃ j ̃] |
||||||
Tap |
Occurrences: 60,154 Examples: * tripé: [t ɾ i p ɛ] * mitre: [m i t ɾ ɨ] * força: [f o ɾ s a] * libré: [l i b ɾ ɛ] |
|||||||
Lateral |
Occurrences: 27,444 Examples: * folfe: [f ɔ l f ɨ] * alces: [ɐ l k ʃ] * arlen: [a x l e ̃] * lilia: [l i l j ɐ] |
Occurrences: 4,831 Examples: * talis: [t a ʎ i s] * lhano: [ʎ ɐ n u] * libas: [ʎ i b ɐ s] * lee: [ʎ i] |
Vowels#
Vowel symbols to the left of are unrounded and those to the right are rounded.
Oral Vowels#
Front |
Near-Front |
Central |
Near-Back |
Back |
|
---|---|---|---|---|---|
Close |
Occurrences: 72,686 Examples: * elrey: [i l ʁ ɛ i] * ricos: [x i k u s] * raiar: [x a i a x] * henry: [e ̃ x i] |
Occurrences: 34,357 Examples: * penfe: [p e ̃ f ɨ] * necio: [n ɨ s j u] * stern: [s t ɨ ɾ n] * sno: [ɛ s ɨ ɛ n ɨ ɔ] |
Occurrences: 63,802 Examples: * odos: [o ð u ʃ] * couro: [k o w ɾ u] * nugas: [n u ɣ ɐ ʃ] * avizo: [ɐ v i z u] |
||
Close-Mid |
Occurrences: 24,251 Examples: * ray: [x a e j] * eixo: [e j ʃ u] * evita: [e v i t ɐ] * eira: [e j ɾ a] |
Occurrences: 20,181 Examples: * gigô: [ʒ i ɡ o] * fôro: [f o ɾ u] * errou: [e x o w] * sopro: [s o p ɾ u] |
|||
Open-Mid |
Occurrences: 11,925 Examples: * orfel: [ɔ ɾ f ɛ l] * fémm: [f ɛ m] * annel: [ɐ ̃ n ɛ l] * srta: [ɛ s i ɛ x i t ɐ] |
Occurrences: 8,994 Examples: * solda: [s ɔ w d ɐ] * onora: [o n ɔ ɾ ɐ] * orva: [ɔ ɾ v ɐ] * roças: [ʁ ɔ s ɐ ʃ] |
|||
Occurrences: 72,022 Examples: * emaús: [i m ɐ u ʃ] * abib: [ɐ β i b] * pisas: [p ɨ z ɐ ʃ] * apega: [ɐ p ɛ ɣ ɐ] |
|||||
Open |
Occurrences: 60,336 Examples: * ápice: [a p i s i] * bafo: [b a f u] * tao: [t a w] * ahia: [a j ɐ] |
Nasal Vowels#
Front |
Near-Front |
Central |
Near-Back |
Back |
|
---|---|---|---|---|---|
Close |
Occurrences: 8,173 Examples: * info: [i ̃ f u] * tinja: [t i ̃ ʒ ɐ] * munim: [m u n i ̃] * main: [m ɐ i ̃] |
Occurrences: 1,756 Examples: * junta: [ʒ u ̃ t ɐ] * com: [k u ̃] * summa: [s u ̃ m ɐ] * lund: [l u ̃ d] |
|||
Close-Mid |
Occurrences: 14,567 Examples: * sem: [s e ̃ j ̃] * lemb: [l e ̃ b] * movem: [m o v e ̃ j ̃] * rennó: [x e ̃ n ɔ] |
Occurrences: 8,439 Examples: * hinom: [i ̃ o ̃] * órion: [ɔ ɾ j o ̃] * maõs: [m o ̃ j ̃ ʃ] * tombo: [t o ̃ b u] |
|||
Open-Mid |
|||||
Occurrences: 21,841 Examples: * negam: [n e ɡ ɐ ̃ w ̃] * imã: [i m ɐ ̃] * pião: [p j ɐ ̃ w ̃] * rem: [ʁ ɐ ̃ j ̃] |
|||||
Open |