Spanish (Spain) MFA dictionary v2.0.0a#

@techreport{mfa_spanish_spain_mfa_dictionary_2022,
	author={McAuliffe, Michael and Sonderegger, Morgan},
	title={Spanish (Spain) MFA dictionary v2.0.0a},
	address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Spanish/Spanish (Spain) MFA dictionary v2_0_0a.html}},
	year={2022},
	month={May},
}
../../_images/full_logo_yellow.svg

Installation#

Install from the MFA command line:

mfa model download dictionary spanish_spain_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 Spanish transcripts.

This dictionary uses the MFA phone set for Spanish, and was used in training the Spanish 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

Nasal

Occurrences:
22,296
Examples:
* tomar:
[t ̪ o m a ɾ]
* maní:
[m a n i]
* muela:
[m w e l a]
* mason:
[m a s o n]
Occurrences:
37,632
Examples:
* andré:
[a n d ̪ ɾ e]
* única:
[u n i k a]
* bloom:
[b l u n]
* mine:
[m i n e]
Occurrences:
2,400
Examples:
* sueña:
[s w e ɲ a]
* paño:
[p a ɲ o]
* pañil:
[p a ɲ i l]
* goñi:
[ɡ o ɲ i]
Occurrences:
2,615
Examples:
* yunco:
[ɟ ʝ u ŋ k o]
* fango:
[f a ŋ ɡ o]
* ing:
[i ŋ]
* vengo:
[b e ŋ ɡ o]

Stop

Occurrences:
16,350
Examples:
* pasó:
[p a s o]
* pego:
[p e ɣ o]
* síp:
[s i p]
* capas:
[k a p a s]
Occurrences:
6,900
Examples:
* braun:
[b ɾ a u n]
* batió:
[b a t ̪ j o]
* borra:
[b o r a]
* botín:
[b o t ̪ i n]
Occurrences:
35,977
Examples:
* cauta:
[k a u t ̪ a]
* yates:
[ɟ ʝ a t ̪ e s]
* monta:
[m o n t ̪ a]
* batán:
[b a t ̪ a n]
Occurrences:
10,206
Examples:
* vende:
[b e n d ̪ e]
* andi:
[a n d ̪ i]
* bland:
[b l a n d ̪]
* dile:
[d ̪ i l e]
Occurrences:
2,331
Examples:
* caqui:
[k a c i]
* equis:
[e c i s]
* akira:
[a c i ɾ a]
* gorki:
[ɡ o r c i]
Occurrences:
194
Examples:
* yale:
[ɟ ʝ a l]
* yurí:
[ɟ ʝ u ɾ i]
* yves:
[ɟ ʝ β e s]
* yesca:
[ɟ ʝ e s k a]
Occurrences:
23,545
Examples:
* naco:
[n a k o]
* sacó:
[s a k o]
* clube:
[k l u β e]
* zueco:
[θ w e k o]
Occurrences:
2,478
Examples:
* goncz:
[ɡ o n t ʃ a s]
* greco:
[ɡ ɾ e k o]
* game:
[ɡ a m e]
* song:
[s o ŋ ɡ]

Affricate

Occurrences:
2,595
Examples:
* chigi:
[t ʃ i ʝ i]
* char:
[t ʃ a ɾ]
* eche:
[e t ʃ e]
* chef:
[t ʃ e f]
Occurrences:
399
Examples:
* yaya:
[ɟ ʝ a ʝ a]
* yuan:
[ɟ ʝ w a n]
* yines:
[ɟ ʝ i n e s]
* yembe:
[ɟ ʝ e m b e]

Sibilant

Occurrences:
54,606
Examples:
* antes:
[a n t ̪ e s]
* samos:
[s a m o s]
* sajá:
[s a x a]
* isler:
[i s l e ɾ]
Occurrences:
280
Examples:
* hacha:
[a t ʃ a]
* chivo:
[t ʃ i β o]
* smash:
[s m a ʃ]
* chapa:
[t ʃ a p a]

Fricative

Occurrences:
7,766
Examples:
* fire:
[f a ʝ e ɾ]
* fiada:
[f j a ð a]
* furia:
[f u ɾ j a]
* draft:
[d ̪ ɾ a f t ̪]
Occurrences:
14,861
Examples:
* cindy:
[θ i n d ̪ i]
* acial:
[a θ j a l]
* razor:
[r a θ o ɾ]
* decía:
[d ̪ e θ i a]
Occurrences:
33,090
Examples:
* acid:
[a θ i ð]
* odre:
[o ð ɾ e]
* budín:
[b u ð i n]
* adán:
[a ð a n]
Occurrences:
3,138
Examples:
* georg:
[ç e o r ɣ]
* jemer:
[ç e m e ɾ]
* meiji:
[m e i ç i]
* argel:
[a r ç e l]
Occurrences:
1,271
Examples:
* yak:
[ɟ ʝ a k]
* iowa:
[ɟ ʝ o w a]
* shogi:
[s o ʝ i]
* yines:
[ɟ ʝ i n e s]

Approximant

Occurrences:
3,594
Examples:
* padua:
[p a ð w a]
* fuese:
[f w e s e]
* pues:
[p w e s]
* wey:
[w e i]
Occurrences:
12,012
Examples:
* lucie:
[l u θ j e]
* rubio:
[r u β j o]
* osmio:
[o s m j o]
* katie:
[k a t ̪ j e]

Tap

Occurrences:
37,943
Examples:
* siria:
[s i ɾ j a]
* aimar:
[a i m a ɾ]
* sidra:
[s i ð ɾ a]
* ungir:
[u ɲ ç i ɾ]

Trill

Occurrences:
17,275
Examples:
* parta:
[p a r t ̪ a]
* avery:
[a β e r i]
* cerda:
[θ e r ð a]
* arpón:
[a r p o n]

Lateral

Occurrences:
24,810
Examples:
* molar:
[m o l a ɾ]
* bucle:
[b u k l e]
* bolis:
[b o l i s]
* herzl:
[e r θ l]
Occurrences:
2,149
Examples:
* pulla:
[p u ʎ a]
* ellen:
[e ʎ e n]
* lloyd:
[ʎ o ʝ ð]
* bell:
[b e ʎ]

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:
48,462
Examples:
* play:
[p l e i]
* salí:
[s a l i]
* pis:
[p i s]
* estío:
[e s t ̪ i o]
Occurrences:
15,890
Examples:
* pude:
[p u ð e]
* out:
[a u t ̪]
* zurda:
[θ u r ð a]
* agudo:
[a ɣ u ð o]

Close-Mid

Occurrences:
65,816
Examples:
* estoy:
[e s t ̪ o i]
* equis:
[e c i s]
* tests:
[t ̪ e s t ̪ s]
* late:
[l a t ̪ e]
Occurrences:
66,612
Examples:
* bond:
[b o n d ̪]
* zuko:
[θ u k o]
* aliso:
[a l i s o]
* hoces:
[o θ e s]

Open-Mid

Open

Occurrences:
113,987
Examples:
* cuita:
[k w i t ̪ a]
* reach:
[r e a k]
* saeta:
[s a e t ̪ a]
* coña:
[k o ɲ a]