Spanish MFA dictionary v2.0.0#

Maintainer: Montreal Forced Aligner
Language: Spanish
Dialect: N/A
Phone set: MFA
Number of words: 86,746
Phones: a b c d̪ e f i j k l m n o p r s tʃ t̪ u w x ç ð ŋ ɟ ɟʝ ɡ ɣ ɲ ɾ ʃ ʎ ʝ β θ
License: CC BY 4.0
Compatible MFA version: v2.0.0
Citation:

@techreport{mfa_spanish_mfa_dictionary_2022,
	author={McAuliffe, Michael and Sonderegger, Morgan},
	title={Spanish MFA dictionary v2.0.0},
	address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Spanish/Spanish MFA dictionary v2_0_0.html}},
	year={2022},
	month={Mar},
}

If you have comments or questions about this dictionary or its phone set, you can check previous MFA model discussion posts or create a new one.

Acoustic models

Spanish MFA acoustic model v2.0.0

Installation#

Install from the MFA command line:

mfa model download dictionary spanish_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	m Occurrences: 24,314 Examples: * monta: [m o n t̪ a] * gemma: [ç e m a] * moe: [m o e] * model: [m o ð e l]			n Occurrences: 43,181 Examples: * green: [ɡ ɾ e e n] * hayan: [a ʝ a n] * hunda: [u n d̪ a] * bone: [b o n e]		ɲ Occurrences: 2,589 Examples: * tiña: [t̪ i ɲ a] * cañar: [k a ɲ a ɾ] * niega: [ɲ j e ɣ a] * ñame: [ɲ a m e]	ŋ Occurrences: 2,827 Examples: * bingo: [b i ŋ ɡ o] * tenca: [t̪ e ŋ k a] * ing: [i ŋ] * tengo: [t̪ e ŋ ɡ o]
Stop	p Occurrences: 18,168 Examples: * hype: [ç i p e] * sopa: [s o p a] * espín: [e s p i n] * apero: [a p e ɾ o] b Occurrences: 7,643 Examples: * bone: [b o n e] * bingo: [b i ŋ ɡ o] * vagón: [b a ɣ o n] * vivo: [b i β o]		t̪ Occurrences: 39,054 Examples: * atea: [a t̪ e a] * hotel: [o t̪ e l] * monta: [m o n t̪ a] * atril: [a t̪ ɾ i l] d̪ Occurrences: 11,351 Examples: * hunda: [u n d̪ a] * wanda: [w a n d̪ a] * debió: [d̪ e β j o] * dura: [d̪ u ɾ a]			c Occurrences: 2,600 Examples: * queso: [c e s o] * kebab: [c e β a β] * quede: [c e ð e] * kilos: [c i l o s] ɟ Occurrences: 213 Examples: * guido: [ɟ i ð o] * guías: [ɟ i a s] * guisa: [ɟ i s a] * guía: [ɟ i a]	k Occurrences: 25,646 Examples: * sox: [s o k s] * kvas: [k β a s] * loca: [l o k a] * rasca: [r a s k a] ɡ Occurrences: 2,710 Examples: * green: [ɡ ɾ e e n] * bingo: [b i ŋ ɡ o] * golf: [ɡ o l f] * omc: [o m ɡ]
Affricate					tʃ Occurrences: 2,769 Examples: * chino: [tʃ i n o] * achín: [a tʃ i n] * cloch: [k l o tʃ] * chigi: [tʃ i ʝ i]	ɟʝ Occurrences: 570 Examples: * yola: [ɟʝ o l a] * yagüe: [ɟʝ a ɣ w e] * yak: [ɟʝ a k] * jet: [ɟʝ e t̪]
Sibilant				s Occurrences: 64,275 Examples: * ries: [r j e s] * hules: [u l e s] * césar: [s e s a ɾ] * sox: [s o k s]	ʃ Occurrences: 360 Examples: * shawn: [ʃ a n] * idish: [i ð i ʃ] * bajux: [b a x u ʃ] * shore: [ʃ o ɾ e]
Fricative		f Occurrences: 8,541 Examples: * flojo: [f l o x o] * ferri: [f e r i] * golf: [ɡ o l f] * fort: [f o r t̪]	θ Occurrences: 14,848 Examples: * césar: [θ e s a ɾ] * cercó: [θ e r k o] * lucen: [l u θ e n] * ceará: [θ e a ɾ a] ð Occurrences: 34,784 Examples: * ready: [r e a ð i] * oídio: [o i ð j o] * adiar: [a ð j a ɾ] * model: [m o ð e l]			ç Occurrences: 3,413 Examples: * gilí: [ç i l i] * hype: [ç i p e] * gemma: [ç e m a] * geo: [ç e o] ʝ Occurrences: 2,081 Examples: * hayan: [a ʝ a n] * traia: [t̪ ɾ a ʝ a] * buyei: [b u ʝ e i] * ayudó: [a ʝ u ð o]
Approximant	w Occurrences: 3,951 Examples: * wayne: [w a i n e] * wanda: [w a n d̪ a] * wicca: [w i k a] * hawái: [x a w a i]					j Occurrences: 14,659 Examples: * ries: [r j e s] * oídio: [o i ð j o] * adiar: [a ð j a ɾ] * debió: [d̪ e β j o]
Tap				ɾ Occurrences: 41,316 Examples: * green: [ɡ ɾ e e n] * adiar: [a ð j a ɾ] * césar: [s e s a ɾ] * atril: [a t̪ ɾ i l]
Trill				r Occurrences: 19,390 Examples: * ries: [r j e s] * ready: [r e a ð i] * arrow: [a r o u] * ferri: [f e r i]
Lateral				l Occurrences: 27,378 Examples: * gilí: [ç i l i] * laho: [l a o] * hules: [u l e s] * yola: [ɟʝ o l a]		ʎ Occurrences: 2,148 Examples: * allah: [a ʎ a] * tallo: [t̪ a ʎ o] * hello: [e ʎ o] * belle: [b e ʎ e]

Vowels#

Vowel symbols to the left of are unrounded and those to the right are rounded.

	Front	Central	Back
Close	i Occurrences: 53,822 Examples: * gilí: [ç i l i] * ready: [r e a ð i] * oídio: [o i ð j o] * hype: [ç i p e]		u Occurrences: 17,452 Examples: * hules: [u l e s] * hunda: [u n d̪ a] * arrow: [a r o u] * sutil: [s u t̪ i l]

Close-Mid	e Occurrences: 74,430 Examples: * atea: [a t̪ e a] * ries: [r j e s] * green: [ɡ ɾ e e n] * ready: [r e a ð i]		o Occurrences: 73,294 Examples: * oídio: [o i ð j o] * laho: [l a o] * bone: [b o n e] * bingo: [b i ŋ ɡ o]

Open-Mid

Open		a Occurrences: 123,922 Examples: * atea: [a t̪ e a] * ready: [r e a ð i] * laho: [l a o] * hayan: [a ʝ a n]