Ctrl+K
Ctrl+K

Portuguese (Portugal) MFA dictionary v2.0.0#

@techreport{mfa_portuguese_portugal_mfa_dictionary_2022,
	author={McAuliffe, Michael and Sonderegger, Morgan},
	title={Portuguese (Portugal) MFA dictionary v2.0.0},
	address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Portuguese/Portuguese (Portugal) MFA dictionary v2_0_0.html}},
	year={2022},
	month={Mar},
}

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

Installation#

Install from the MFA command line:

mfa model download dictionary portuguese_portugal_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:
19,597
Examples:
* anama:
[ɐ n ɐ m ɐ]
* hmm:
[m]
* jarmé:
[ʒ ɐ ɾ m ɛ]
* húmus:
[u m u ʃ]
Occurrences:
13,264
Examples:
* núas:
[n u ɐ ʃ]
* anama:
[ɐ n ɐ m ɐ]
* vigny:
[v i ɣ ɨ n i]
* lana:
[l ɐ n ɐ]
Occurrences:
2,827
Examples:
* cunha:
[k u ɲ ɐ]
* unhão:
[u ɲ ɐ̃ w̃]
* vinho:
[v i ɲ u]
* ganhá:
[ɡ ɐ ɲ a]

Stop
Occurrences:
19,255
Examples:
* opõem:
[ɔ p õ j̃ ɐ̃ j̃]
* cipó:
[s i p ɔ]
* paria:
[p ɐ ɾ j ɐ]
* paty:
[p a tʃ i]
Occurrences:
6,637
Examples:
* baía:
[b ɐ i ɐ]
* abuso:
[ɐ b u z u]
* fubir:
[f u b i ɾ]
* bater:
[b ɐ t e ɾ]
Occurrences:
36,727
Examples:
* satan:
[s ɐ t ɐ̃]
* fitos:
[f i t u ʃ]
* feltz:
[f ɛ l t ʃ]
* tugue:
[t u ɣ ɨ]
Occurrences:
16,539
Examples:
* aindu:
[ɐ ĩ d u]
* drená:
[d ɾ e n a]
* pendi:
[p ẽ d i]
* uandu:
[w ɐ̃ d u]
Occurrences:
1,505
Examples:
* etc:
[i t c i]
* inesc:
[i n ɨ ʃ c i]
* kiko:
[c i k u]
* quês:
[c e ʃ]
Occurrences:
287
Examples:
* guido:
[ɟ i ð u]
* guita:
[ɟ i t ɐ]
* eguer:
[e ɟ e ɾ]
* gps:
[ɟ e p e ɛ s ɨ]
Occurrences:
24,026
Examples:
* corar:
[k ɔ ɾ a ɾ]
* couce:
[k o w s ɨ]
* coifa:
[k w i f ɐ]
* cupom:
[k u p õ]
Occurrences:
4,397
Examples:
* games:
[ɡ ɐ m ɨ ʃ]
* gosam:
[ɡ o z ɐ̃ w̃]
* magno:
[m a ɡ n u]
* gabo:
[ɡ a b u]

Affricate
Occurrences:
814
Examples:
* paty:
[p a tʃ i]
* tweet:
[t w i tʃ i]
* start:
[s t a ʁ tʃ i]
* vete:
[v ɛ tʃ i]
Occurrences:
510
Examples:
* gude:
[ɡ u dʒ i]
* deos:
[dʒ i u ʃ]
* dita:
[dʒ i t ɐ]
* diego:
[dʒ i e ɣ u]

Sibilant
Occurrences:
26,142
Examples:
* satan:
[s ɐ t ɐ̃]
* cipó:
[s i p ɔ]
* couce:
[k o w s ɨ]
* cedia:
[s ɨ ð i ɐ]
Occurrences:
9,542
Examples:
* abuso:
[ɐ b u z u]
* mouse:
[m o w z i]
* acuse:
[ɐ k u z i]
* zomba:
[z õ b ɐ]
Occurrences:
36,323
Examples:
* núas:
[n u ɐ ʃ]
* fitos:
[f i t u ʃ]
* feltz:
[f ɛ l t ʃ]
* expôr:
[ɐ j ʃ p o ɾ]
Occurrences:
6,526
Examples:
* jarmé:
[ʒ ɐ ɾ m ɛ]
* jaca:
[ʒ ɐ k ɐ]
* jan:
[ʒ ɐ̃]
* giro:
[ʒ i ʁ u]

Fricative
Occurrences:
10,334
Examples:
* fitos:
[f i t u ʃ]
* rufam:
[ʁ u f ɐ̃ w̃]
* feltz:
[f ɛ l t ʃ]
* coifa:
[k w i f ɐ]
Occurrences:
13,838
Examples:
* vigny:
[v i ɣ ɨ n i]
* uva:
[u v ɐ]
* evite:
[i v i t ɨ]
* cerva:
[s ɨ ɾ v ɐ]
Occurrences:
16,618
Examples:
* leda:
[l ɨ ð ɐ]
* cedia:
[s ɨ ð i ɐ]
* redor:
[ʁ ɨ ð o ɾ]
* codó:
[k ɔ ð ɔ]
Occurrences:
9,875
Examples:
* rufam:
[ʁ u f ɐ̃ w̃]
* redor:
[ʁ ɨ ð o ɾ]
* reate:
[ʁ j a t ɨ]
* terra:
[t ɛ ʁ ɐ]

Approximant
Occurrences:
8,424
Examples:
* couce:
[k o w s ɨ]
* legou:
[l ɨ ɣ o w]
* coifa:
[k w i f ɐ]
* mouse:
[m o w z i]
Occurrences:
6,605
Examples:
* rufam:
[ʁ u f ɐ̃ w̃]
* terão:
[t ɨ ɾ ɐ̃ w̃]
* gosam:
[ɡ o z ɐ̃ w̃]
* talão:
[t ɐ l ɐ̃ w̃]
Occurrences:
17,646
Examples:
* paria:
[p ɐ ɾ j ɐ]
* lear:
[l j a ɾ]
* expôr:
[ɐ j ʃ p o ɾ]
* noite:
[n o j t ɨ]
Occurrences:
3,214
Examples:
* opõem:
[ɔ p õ j̃ ɐ̃ j̃]
* mãi:
[m ɐ̃ j̃]
* tomem:
[t ɔ m ẽ j̃]
* qtiem:
[k tʃ ĩ ɐ̃ j̃]

Tap
Occurrences:
48,730
Examples:
* corar:
[k ɔ ɾ a ɾ]
* paria:
[p ɐ ɾ j ɐ]
* lear:
[l j a ɾ]
* expôr:
[ɐ j ʃ p o ɾ]

Lateral
Occurrences:
23,441
Examples:
* leda:
[l ɨ ð ɐ]
* feltz:
[f ɛ l t ʃ]
* legou:
[l ɨ ɣ o w]
* lana:
[l ɐ n ɐ]
Occurrences:
1,999
Examples:
* calha:
[k ɐ ʎ ɐ]
* olhos:
[ɔ ʎ u ʃ]
* colha:
[k u ʎ ɐ]
* elhor:
[ɨ ʎ ɔ ɾ]

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:
50,627
Examples:
* cipó:
[s i p ɔ]
* fitos:
[f i t u ʃ]
* baía:
[b ɐ i ɐ]
* vigny:
[v i ɣ ɨ n i]
Occurrences:
34,357
Examples:
* couce:
[k o w s ɨ]
* leda:
[l ɨ ð ɐ]
* vigny:
[v i ɣ ɨ n i]
* legou:
[l ɨ ɣ o w]
Occurrences:
51,438
Examples:
* aindu:
[ɐ ĩ d u]
* núas:
[n u ɐ ʃ]
* fitos:
[f i t u ʃ]
* rufam:
[ʁ u f ɐ̃ w̃]

Close-Mid
Occurrences:
10,571
Examples:
* bater:
[b ɐ t e ɾ]
* elena:
[i l e n ɐ]
* drená:
[d ɾ e n a]
* lêem:
[l e ɐ̃ j]
Occurrences:
11,364
Examples:
* couce:
[k o w s ɨ]
* legou:
[l ɨ ɣ o w]
* expôr:
[ɐ j ʃ p o ɾ]
* redor:
[ʁ ɨ ð o ɾ]

Open-Mid
Occurrences:
9,111
Examples:
* feltz:
[f ɛ l t ʃ]
* jarmé:
[ʒ ɐ ɾ m ɛ]
* terra:
[t ɛ ʁ ɐ]
* lerna:
[l ɛ ɾ n ɐ]
Occurrences:
7,134
Examples:
* opõem:
[ɔ p õ j̃ ɐ̃ j̃]
* corar:
[k ɔ ɾ a ɾ]
* cipó:
[s i p ɔ]
* horn:
[ɔ ɾ n]
Occurrences:
63,312
Examples:
* aindu:
[ɐ ĩ d u]
* satan:
[s ɐ t ɐ̃]
* núas:
[n u ɐ ʃ]
* paria:
[p ɐ ɾ j ɐ]

Open
Occurrences:
37,775
Examples:
* corar:
[k ɔ ɾ a ɾ]
* paty:
[p a tʃ i]
* lear:
[l j a ɾ]
* reate:
[ʁ j a t ɨ]

Nasal Vowels#

Front

Near-Front

Central

Near-Back

Back

Close
Occurrences:
5,648
Examples:
* aindu:
[ɐ ĩ d u]
* insta:
[ĩ ʃ t ɐ]
* impos:
[ĩ p u ʃ]
* qtiem:
[k tʃ ĩ ɐ̃ j̃]
Occurrences:
1,326
Examples:
* summo:
[s ũ m u]
* jungi:
[ʒ ũ ʒ i]
* anuns:
[ɐ n ũ ʃ]
* unçaõ:
[ũ s õ]

Close-Mid
Occurrences:
10,878
Examples:
* pendi:
[p ẽ d i]
* tomem:
[t ɔ m ẽ j̃]
* tenta:
[t ẽ t ɐ]
* enviá:
[ẽ v j a]
Occurrences:
6,266
Examples:
* opõem:
[ɔ p õ j̃ ɐ̃ j̃]
* cupom:
[k u p õ]
* zomba:
[z õ b ɐ]
* contê:
[k õ t e]

Open-Mid
Occurrences:
17,132
Examples:
* satan:
[s ɐ t ɐ̃]
* opõem:
[ɔ p õ j̃ ɐ̃ j̃]
* rufam:
[ʁ u f ɐ̃ w̃]
* terão:
[t ɨ ɾ ɐ̃ w̃]

Open
On this page
Show Source