Guarani CV dictionary v2.0.0#

Maintainer: Vox Communis
Language: Guarani
Dialect: N/A
Phone set: XPF
Number of words: 1,103
Phones: a dʒ e i k l o p r s t u x ã õ ĩ ũ ɯ ɰ ɾ ʃ ʋ ʔ ̃ ᵐb ᵑɡ ẽ ⁿd
License: CC-0
Compatible MFA version: v2.0.0
Citation:

@misc{Ahn_Chodroff_2022,
	author={Ahn, Emily and Chodroff, Eleanor},
	title={VoxCommunis Corpus},
	address={\url{https://osf.io/t957v}},
	publisher={OSF},
	year={2022},
	month={Jan}
}

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

Guarani CV acoustic model v2.0.0

Installation#

Install from the MFA command line:

mfa model download dictionary guarani_cv

Or download from the release page.

Intended use#

This dictionary is intended for forced alignment of Guarani transcripts.

This dictionary uses the XPF phone set for Guarani, and was used in training the Guarani XPF 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	Alveolar	Alveopalatal	Velar	Glottal
Stop	p Occurrences: 430 Examples: * opu'ã: [o p u ʔ ã] * moõpa: [ᵐb o õ p a] * epoi: [e p o i] * upe: [u p e]		t Occurrences: 281 Examples: * añete: [a dʒ e t e] * pizza: [p i t s a] * tata: [t a t a] * toro: [t o ɾ o]		k Occurrences: 349 Examples: * ka'a: [k a ʔ a] * aníke: [a ⁿd i k e] * ka'ay: [k a ʔ a ɯ] * ojuka: [o dʒ u k a]	ʔ Occurrences: 259 Examples: * opu'ã: [o p u ʔ ã] * ru'ã: [ɾ u ʔ ã] * ka'a: [k a ʔ a] * ka'ay: [k a ʔ a ɯ]
Affricate				dʒ Occurrences: 302 Examples: * añete: [a dʒ e t e] * ñana: [dʒ a ⁿd a] * ñande: [dʒ a ⁿd e] * oñeno: [o dʒ e ⁿd o]
Sibilant			s Occurrences: 120 Examples: * vosa: [ʋ o s a] * ohasa: [o x a s a] * osẽva: [o s ẽ ʋ a] * pizza: [p i t s a]	ʃ Occurrences: 67 Examples: * michĩ: [ᵐb i ʃ ĩ] * luchi: [l u ʃ i] * chupe: [ʃ u p e] * chipa: [ʃ i p a]
Approximant		ʋ Occurrences: 283 Examples: * vosa: [ʋ o s a] * oĩva: [o ĩ ʋ a] * osẽva: [o s ẽ ʋ a] * ndéve: [ⁿd e ʋ e]			ɰ Occurrences: 128 Examples: * ogami: [o ɰ a ᵐb i] * hogue: [x o ɰ u e] * jagua: [dʒ a ɰ u a] * guyra: [ɰ u ɯ ɾ a]
Tap			ɾ Occurrences: 517 Examples: * urdu: [u ɾ ⁿd u] * are: [a ɾ e] * ru'ã: [ɾ u ʔ ã] * ryvy: [ɾ ɯ ʋ ɯ]
Trill			r Occurrences: 2 Examples:
Lateral			l Occurrences: 31 Examples: * luchi: [l u ʃ i] * clari: [k l a ɾ i] * helen: [x e l e ⁿd] * leõ: [l e õ]

Vowels#

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

Oral Vowels#

	Front	Central	Back
Close	i Occurrences: 578 Examples: * epoi: [e p o i] * michĩ: [ᵐb i ʃ ĩ] * maína: [ᵐb a i ⁿd a] * luchi: [l u ʃ i]		ɯ Occurrences: 311 Examples: * ijyvy: [i dʒ ɯ ʋ ɯ] * ryvy: [ɾ ɯ ʋ ɯ] * ka'ay: [k a ʔ a ɯ] * vy'a: [ʋ ɯ ʔ a] u Occurrences: 487 Examples: * opu'ã: [o p u ʔ ã] * urdu: [u ɾ ⁿd u] * upe: [u p e] * ru'ã: [ɾ u ʔ ã]

Close-Mid	e Occurrences: 1,039 Examples: * epoi: [e p o i] * upe: [u p e] * are: [a ɾ e] * añete: [a dʒ e t e]		o Occurrences: 738 Examples: * opu'ã: [o p u ʔ ã] * vosa: [ʋ o s a] * moõpa: [ᵐb o õ p a] * epoi: [e p o i]

Open-Mid

Open		a Occurrences: 1,255 Examples: * vosa: [ʋ o s a] * moõpa: [ᵐb o õ p a] * are: [a ɾ e] * ka'a: [k a ʔ a]

Nasal Vowels#

	Front	Central	Back
Close	ĩ Occurrences: 16 Examples: * michĩ: [ᵐb i ʃ ĩ] * oĩva: [o ĩ ʋ a] * omoĩ: [o ᵐb o ĩ] * peteĩ: [p e t e ĩ]		ũ Occurrences: 21 Examples: * irũ: [i ɾ ũ] * hũngy: [x ũ ᵑɡ ɯ]

Close-Mid	ẽ Occurrences: 49 Examples: * osẽva: [o s ẽ ʋ a] * iko'ẽ: [i k o ʔ ẽ] * he'ẽ: [x e ʔ ẽ] * osẽ: [o s ẽ]		õ Occurrences: 43 Examples: * moõpa: [ᵐb o õ p a] * leõ: [l e õ] * ñavõ: [dʒ a ʋ õ] * pokõi: [p o k õ i]

Open-Mid

Open		ã Occurrences: 100 Examples: * opu'ã: [o p u ʔ ã] * ru'ã: [ɾ u ʔ ã] * omosã: [o ᵐb o s ã] * tupã: [t u p ã]