Korean (Jamo) MFA dictionary v2.0.0#

Maintainer: Montreal Forced Aligner
Language: Korean
Dialect: N/A
Phone set: MFA
Number of words: 54,093
Phones: b d dʑ e eː h i iː j k kʰ k̚ k͈ m n o oː p pʰ p̚ p͈ s sʰ s͈ t tɕ tɕʰ tɕ͈ tʰ t̚ t͈ u uː w x ç ŋ ɐ ɕʰ ɕ͈ ɡ ɣ ɥ ɦ ɨ ɨː ɭ ɰ ɲ ɸ ɾ ʌ ʌː ʎ ʝ β
License: CC BY 4.0
Compatible MFA version: v2.0.0
Citation:

@techreport{mfa_korean_jamo_mfa_dictionary_2022,
	author={McAuliffe, Michael and Sonderegger, Morgan},
	title={Korean (Jamo) MFA dictionary v2.0.0},
	address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Korean/Korean (Jamo) 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.

G2P models

Korean (Jamo) MFA G2P model v2.0.0

Installation#

Install from the MFA command line:

mfa model download dictionary korean_jamo_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 Korean transcripts.

This dictionary uses the MFA phone set for Korean, and was used in training the Korean 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	Alveolar	Retroflex	Palatal	Velar	Glottal
Nasal	m Occurrences: 15,143 Examples: * ㅁㅗㄷㅜ: [m o d u] * ㅁㅗㅍㅣ: [m o pʰ i] * ㅁㅣㄴㅇㅛ: [m i ɲ o] * ㄴㅏㅁㅁㅐ: [n ɐ m m eː]	n Occurrences: 28,729 Examples: * ㄴㅗㅇㅈㅣ: [n o ŋ dʑ i] * ㄴㅏㅁㅁㅐ: [n ɐ m m eː] * ㄴㅐㄱㅓ: [n eː ɡ ʌ] * ㄴㅐㄱㅏㄱ: [n eː ɡ ɐ k̚]		ɲ Occurrences: 3,463 Examples: * ㅁㅣㄴㅇㅛ: [m i ɲ o] * ㅋㅗㄴㅑㄱ: [kʰ o ɲ ɐ k̚] * ㅈㅣㄴㅈㅏ: [tɕ i ɲ dʑ ɐ] * ㄴㅏㅁㄴㅕ: [n ɐ m ɲ ʌ]	ŋ Occurrences: 15,801 Examples: * ㄴㅗㅇㅈㅣ: [n o ŋ dʑ i] * ㄸㅏㅇㅇㅣ: [t͈ ɐ ŋ i] * ㅇㅘㅇㅈㅏ: [w ɐ ŋ dʑ ɐ] * ㅂㅣㅇㅅㅜ: [p i ŋ sʰ u]
Stop	p Occurrences: 5,399 Examples: * ㅂㅣㅇㅅㅜ: [p i ŋ sʰ u] * ㅂㅏㅇㅟ: [p ɐ ɥ i] * ㅂㅗㄴㅐㅆ: [p o n eː t̚] * ㅂㅏㄷㅡㄴ: [p ɐ d ɨ n] b Occurrences: 4,167 Examples: * ㅇㅖㅂㅗ: [j e b o] * ㄲㅏㅂㅔㄹ: [k͈ ɐ b e ɭ] * ㅈㅜㄴㅂㅣ: [tɕ uː n b i] * ㅅㅔㅂㅜㄴ: [sʰ e b u n]	t Occurrences: 3,929 Examples: * ㄷㅓㄹㄷㅏ: [t ʌː ɭ d ɐ] * ㄷㅔㄴㅣ: [t e n i] * ㄷㅙㄷㅗ: [t w eː d o] * ㄷㅡㅇㅍㅣ: [t ɨ ŋ pʰ i] d Occurrences: 9,538 Examples: * ㅁㅗㄷㅜ: [m o d u] * ㅈㅝㄷㅗ: [tɕ w ʌ d o] * ㄴㅏㄷㅏㅁ: [n ɐ d ɐ m] * ㅂㅏㄷㅡㄴ: [p ɐ d ɨ n]			k Occurrences: 8,928 Examples: * ㄱㅣㅇㅏ: [k i ɐ] * ㄱㅣㄹㄹㅐ: [k i ɭ ɭ eː] * ㄱㅡㄴㅔ: [k ɨː n e] * ㅅㅗㄱㅅㅔ: [sʰ o k s e] ɡ Occurrences: 12,741 Examples: * ㄴㅐㄱㅓ: [n eː ɡ ʌ] * ㄴㅐㄱㅏㄱ: [n eː ɡ ɐ k̚] * ㅈㅓㄱㅡㅁ: [tɕ ʌː ɡ ɨ m] * ㅁㅏㄱㅜ: [m ɐ ɡ u]
Affricate				tɕ Occurrences: 6,425 Examples: * ㅈㅓㄱㅡㅁ: [tɕ ʌː ɡ ɨ m] * ㅈㅝㄷㅗ: [tɕ w ʌ d o] * ㅈㅣㄴㅈㅏ: [tɕ i ɲ dʑ ɐ] * ㅈㅜㄴㅂㅣ: [tɕ uː n b i] dʑ Occurrences: 8,444 Examples: * ㄴㅗㅇㅈㅣ: [n o ŋ dʑ i] * ㅇㅘㅇㅈㅏ: [w ɐ ŋ dʑ ɐ] * ㅃㅏㅈㅣㅁ: [p͈ ɐ dʑ i m] * ㅇㅠㅈㅗ: [j u dʑ o]
Sibilant Plain		s Occurrences: 2,055 Examples: * ㅅㅗㄱㅅㅔ: [sʰ o k s e] * ㅌㅏㄹㅅㅔ: [tʰ ɐ ɭ s e] * ㅅㅓㄹㅅㅏ: [sʰ ʌ ɭ s ɐ] * ㅎㅐㅆㅇㅓ: [h eː s ʌ]
Tense		s͈ Occurrences: 1,094 Examples: * ㅆㅡㄹㅜ: [s͈ ɨ ɾ u] * ㅆㅡㅍㅜ: [s͈ ɨ pʰ u] * ㅇㅝㄴㅆㅜ: [w ʌ n s͈ u] * ㅆㅏㅇㅁㅜ: [s͈ ɐ ŋ m u]		ɕ͈ Occurrences: 765 Examples: * ㅂㅜㄹㅆㅣ: [p uː ɭ ɕ͈ i] * ㅅㅣㅇㅔ: [ɕ͈ i e] * ㅁㅗㄳㅇㅣ: [m o k ɕ͈ i] * ㅁㅜㄴㅆㅣ: [m uː ɭ ɕ͈ i]
Aspirated		sʰ Occurrences: 11,829 Examples: * ㅂㅣㅇㅅㅜ: [p i ŋ sʰ u] * ㅎㅜㅅㅏㅁ: [ɸ uː sʰ ɐ m] * ㅅㅏㅅㅡㄹ: [sʰ ɐ sʰ ɨ ɭ] * ㅊㅜㅅㅜ: [tɕʰ u sʰ u]		ɕʰ Occurrences: 3,428 Examples: * ㅅㅣㅅㅜㄹ: [ɕʰ i sʰ u ɭ] * ㅅㅣㄱㅂㅣ: [ɕʰ i k̚ pʰ i] * ㅁㅏㅅㅣ: [m ɐ ɕʰ i] * ㅅㅣㅊㅏ: [ɕʰ iː tɕʰ ɐ]
Fricative	ɸ Occurrences: 1,000 Examples: * ㅎㅜㅅㅏㅁ: [ɸ uː sʰ ɐ m] * ㅎㅗㄹㅗ: [ɸ o ɾ o] * ㅎㅘㅅㅡㅇ: [ɸ w ɐ sʰ ɨ ŋ] * ㅎㅘㅇㅎㅜ: [ɸ w ɐ ŋ β u]			ç Occurrences: 447 Examples: * ㅎㅠㅇㄱㅣ: [ç u ŋ ɡ i] * ㅎㅕㄴㅈㅐ: [ç ʌ ɲ tɕ eː] * ㅎㅕㄴㅈㅣ: [ç ʌː ɲ dʑ i] * ㅎㅢㅎㅘ: [ç i β w ɐ] ʝ Occurrences: 829 Examples: * ㅇㅕㅇㅎㅢ: [j ʌ ŋ ʝ i] * ㄱㅣㅎㅕㅇ: [k i ʝ ʌ ŋ] * ㅅㅣㅎㅖ: [ɕʰ i ʝ e] * ㄱㅏㅎㅐㅆ: [k ɐ ʝ eː t̚]		h Occurrences: 4,340 Examples: * ㅌㅐㄱㅎㅐ: [tʰ eː k h eː] * ㅎㅏㄱㅣㄴ: [h ɐ ɡ i n] * ㅎㅏㅇㅇㅢ: [h ɐ ŋ ɰ i] * ㅎㅟㅂㅏㄹ: [h ɥ i b ɐ ɭ] ɦ Occurrences: 2,079 Examples: * ㅊㅣㅎㅏ: [tɕʰ i ɦ ɐ] * ㅇㅏㅎㅏ: [ɐ ɦ ɐ] * ㄷㅐㅎㅏㄴ: [t eː ɦ ɐ n] * ㄷㅗㅎㅏ: [t o ɦ ɐ]
Approximant	w Occurrences: 7,116 Examples: * ㅇㅘㅇㅈㅏ: [w ɐ ŋ dʑ ɐ] * ㅈㅝㄷㅗ: [tɕ w ʌ d o] * ㅇㅚㅊㅣ: [w eː tɕʰ i] * ㅇㅏㄴㅇㅘ: [ɐ n w ɐ]			j Occurrences: 11,681 Examples: * ㅇㅖㅂㅗ: [j e b o] * ㅇㅖㅎㅜ: [j eː β u] * ㅁㅕㄴㅇㅢ: [m j ʌː n ɰ i] * ㅁㅓㄱㅇㅕ: [m ʌ ɡ j ʌ] ɥ Occurrences: 1,108 Examples: * ㅂㅏㅇㅟ: [p ɐ ɥ i] * ㅅㅏㅇㅟ: [sʰ ɐ ɥ i] * ㄱㅟㅊㅜ: [k ɥ i tɕʰ u] * ㅇㅟㄱㅏ: [ɥ i ɡ ɐ]	ɰ Occurrences: 1,682 Examples: * ㅁㅕㄴㅇㅢ: [m j ʌː n ɰ i] * ㄱㅗㅇㅢ: [k o ɰ i] * ㅇㅢㅁㅜㄴ: [ɰ i m u n] * ㅎㅏㅇㅇㅢ: [h ɐ ŋ ɰ i]
Tap		ɾ Occurrences: 11,736 Examples: * ㅆㅡㄹㅜ: [s͈ ɨ ɾ u] * ㅇㅏㅀㅇㅣ: [ɐ ɾ i] * ㅎㅗㄹㅗ: [ɸ o ɾ o] * ㅇㅜㄹㅣㅁ: [u ɾ i m]
Lateral				ʎ Occurrences: 3,213 Examples: * ㅂㅜㄹㄹㅕ: [p u ʎ ʎ ʌ] * ㅁㅓㄹㄹㅣ: [m ʌː ʎ ʎ i] * ㄱㅘㄴㄹㅣ: [k w ɐ ʎ ʎ i] * ㅅㅣㄹㄴㅐ: [ɕʰ i ʎ ʎ eː]
Lateral Tap			ɭ Occurrences: 10,883 Examples: * ㄱㅣㄹㄹㅐ: [k i ɭ ɭ eː] * ㅅㅏㅅㅡㄹ: [sʰ ɐ sʰ ɨ ɭ] * ㅅㅣㅅㅜㄹ: [ɕʰ i sʰ u ɭ] * ㅇㅓㄹㄷㅏ: [ʌː ɭ d ɐ]

Vowels#

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

	Front	Central	Back
Close	i Occurrences: 27,539 Examples: * ㄴㅗㅇㅈㅣ: [n o ŋ dʑ i] * ㄸㅏㅇㅇㅣ: [t͈ ɐ ŋ i] * ㅁㅗㅍㅣ: [m o pʰ i] * ㅁㅣㄴㅇㅛ: [m i ɲ o] iː Occurrences: 1,730 Examples: * ㅇㅣㅎㅗ: [iː β o] * ㅅㅣㅊㅏ: [ɕʰ iː tɕʰ ɐ] * ㅅㅟㅇㅜㄴ: [sʰ ɥ iː u n] * ㅇㅣㅁㄴㅏ: [iː m n ɐ]	ɨ Occurrences: 16,928 Examples: * ㅆㅡㄹㅜ: [s͈ ɨ ɾ u] * ㅈㅓㄱㅡㅁ: [tɕ ʌː ɡ ɨ m] * ㅇㅓㄴㅡ: [ʌ n ɨ] * ㅅㅏㅅㅡㄹ: [sʰ ɐ sʰ ɨ ɭ] ɨː Occurrences: 233 Examples: * ㄱㅡㄴㅔ: [k ɨː n e] * ㅇㅡㅇㅡㅁ: [ɨː ɨ m] * ㅇㅡㅇㅣㅇ: [ɨː i ŋ] * ㄱㅡㅇㅈㅣ: [k ɨː ŋ dʑ i]	u Occurrences: 13,991 Examples: * ㅆㅡㄹㅜ: [s͈ ɨ ɾ u] * ㅁㅗㄷㅜ: [m o d u] * ㅂㅣㅇㅅㅜ: [p i ŋ sʰ u] * ㅇㅖㅎㅜ: [j eː β u] uː Occurrences: 1,038 Examples: * ㅎㅜㅅㅏㅁ: [ɸ uː sʰ ɐ m] * ㅈㅜㄴㅂㅣ: [tɕ uː n b i] * ㅅㅜㅂㅏㄴ: [sʰ uː b ɐ n] * ㅈㅜㅇㅣㄹ: [tɕ uː i ɭ]

Close-Mid	e Occurrences: 9,056 Examples: * ㅇㅖㅂㅗ: [j e b o] * ㄱㅡㄴㅔ: [k ɨː n e] * ㅅㅗㄱㅅㅔ: [sʰ o k s e] * ㄲㅏㅂㅔㄹ: [k͈ ɐ b e ɭ] eː Occurrences: 8,707 Examples: * ㄴㅏㅁㅁㅐ: [n ɐ m m eː] * ㄴㅐㄱㅓ: [n eː ɡ ʌ] * ㄴㅐㄱㅏㄱ: [n eː ɡ ɐ k̚] * ㅇㅖㅎㅜ: [j eː β u]		o Occurrences: 18,232 Examples: * ㄴㅗㅇㅈㅣ: [n o ŋ dʑ i] * ㅁㅗㄷㅜ: [m o d u] * ㅁㅗㅍㅣ: [m o pʰ i] * ㅁㅣㄴㅇㅛ: [m i ɲ o] oː Occurrences: 1,742 Examples: * ㄱㅛㅍㅕㄴ: [k j oː pʰ j ʌ n] * ㅇㅗㅎㅗ: [oː β o] * ㄱㅗㅇㅈㅏ: [k oː ŋ dʑ ɐ] * ㅇㅗㅇㅎㅗ: [oː ŋ β o]

Open-Mid			ʌ Occurrences: 24,653 Examples: * ㄴㅐㄱㅓ: [n eː ɡ ʌ] * ㅇㅓㄴㅡ: [ʌ n ɨ] * ㅈㅝㄷㅗ: [tɕ w ʌ d o] * ㄸㅓㅁㅏㅌ: [t͈ ʌ m ɐ t̚] ʌː Occurrences: 2,398 Examples: * ㅈㅓㄱㅡㅁ: [tɕ ʌː ɡ ɨ m] * ㅁㅕㄴㅇㅢ: [m j ʌː n ɰ i] * ㅅㅓㄴㄴㅐ: [sʰ ʌː n n eː] * ㅂㅕㄴㅁㅗ: [p j ʌː n m o]
		ɐ Occurrences: 42,211 Examples: * ㄸㅏㅇㅇㅣ: [t͈ ɐ ŋ i] * ㄴㅏㅁㅁㅐ: [n ɐ m m eː] * ㅇㅏㅀㅇㅣ: [ɐ ɾ i] * ㅇㅘㅇㅈㅏ: [w ɐ ŋ dʑ ɐ]
Open