Turkish MFA dictionary v2.0.0a#
@techreport{mfa_turkish_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Turkish MFA dictionary v2.0.0a},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Turkish/Turkish MFA dictionary v2_0_0a.html}},
year={2022},
month={May},
}
G2P models Acoustic models |
Installation#
Install from the MFA command line:
mfa model download dictionary turkish_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 Turkish transcripts.
This dictionary uses the MFA phone set for Turkish, and was used in training the Turkish 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 |
Glottal |
---|---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 17,592 Examples: * içmem: [i t ʃ m e m] * ekme: [e c m ɛ] * mudur: [m u d ̪ u ɾ] * emmim: [e m ː i m] Occurrences: 51 Examples: * emmiş: [e m ː i ʃ] * amma: [a m ː a] * jimmy: [ʒ i m ː i] * hımm: [h ɯ m ː] |
Occurrences: 27,022 Examples: * burun: [b u ɾ u n ̪] * izni: [i z ̪ n ̪ ɪ] * önder: [ø n ̪ d ̪ e ɾ] * yaşın: [j a ʃ ɯ n ̪] Occurrences: 76 Examples: * anne: [a n ̪ ː ɛ] * ninni: [n ̪ i n ̪ ː ɪ] * annem: [a n ̪ ː e m] * annen: [a n ̪ ː e n ̪] |
Occurrences: 189 Examples: * süngü: [s ̪ y ɲ ɟ ʏ] * engel: [ɛ ɲ ɟ e ʎ] * dink: [d ̪ i ɲ c] * anki: [a ɲ c ɪ] |
Occurrences: 205 Examples: * enka: [ɛ ŋ k a] * inkâr: [i ŋ k a ɾ] * inga: [i ŋ ɡ a] * ahenk: [a ç ɛ ŋ k] |
||||
Stop |
Occurrences: 3,322 Examples: * pkk: [p e c e c ɛ] * topik: [t ̪ o p i c] * sepet: [s ̪ e p e t ̪] * depo: [d ̪ e p o] Occurrences: 6 Examples: * züppe: [z ̪ y p ː ɛ] * kappa: [k a p ː a] * zappa: [z ̪ a p ː a] Occurrences: 6,159 Examples: * başak: [b a ʃ a k] * ezber: [e z ̪ b e ɾ] * çaba: [t ʃ a b a] * berna: [b ɛ ɾ n ̪ a] Occurrences: 22 Examples: * kubbe: [k u b ː ɛ] * tıbbi: [t ̪ ɯ b ː ɪ] * abbas: [a b ː a s ̪] |
Occurrences: 13,355 Examples: * katti: [k a t ̪ ː ɪ] * çatal: [t ʃ a t ̪ a ɫ] * kenti: [c e n ̪ t ̪ ɪ] * tenha: [t ̪ e n ̪ h a] Occurrences: 630 Examples: * yattı: [j a t ̪ ː ɨ] * itti: [i t ̪ ː ɪ] * tattı: [t ̪ a t ̪ ː ɨ] * gatt: [ɡ a t ̪ ː] Occurrences: 16,616 Examples: * duyan: [d ̪ u j a n ̪] * daya: [d ̪ a j a] * dönüp: [d ̪ ø n ̪ y p] * rpde: [ɾ p d ̪ ɛ] Occurrences: 140 Examples: * maddi: [m a d ̪ ː ɪ] * ciddi: [d ʒ i d ̪ ː ɪ] * zıddı: [z ̪ ɯ d ̪ ː ɨ] * cadde: [d ʒ a d ̪ ː ɛ] |
Occurrences: 5,951 Examples: * bekaj: [b e c a ʒ] * öksüz: [ø c s ̪ y z ̪] * kirli: [c i ɾ ʎ ɪ] * küba: [c y b a] Occurrences: 24 Examples: * ökkeş: [ø c ː e ʃ] * takke: [t ̪ a c ː ɛ] * sikke: [s ̪ i c ː ɛ] Occurrences: 3,321 Examples: * dalga: [d ̪ a ɫ ɟ a] * güreş: [ɟ y ɾ e ʃ] * geber: [ɟ e b e ɾ] * gülüp: [ɟ y ʎ y p] |
Occurrences: 11,248 Examples: * büktü: [b y k t ̪ ʏ] * mark: [m a ɾ k] * kolum: [k o ɫ u m] * kulun: [k u ɫ u n ̪] Occurrences: 84 Examples: * okka: [o k ː a] * akka: [a k ː a] * dakka: [d ̪ a k ː a] * hakkı: [h a k ː ɨ] Occurrences: 685 Examples: * lpg: [ʎ e p e ɡ ɛ] * gatt: [ɡ a t ̪ ː] * saygı: [s ̪ a j ɡ ɨ] * damga: [d ̪ a m ɡ a] Occurrences: 2 Examples: |
||||
Affricate |
Occurrences: 3,645 Examples: * çakal: [t ʃ a k a ɫ] * çıkış: [t ʃ ɯ c ɾ ɯ ʃ] * uçan: [u t ʃ a n ̪] * açtı: [a t ʃ t ̪ ɨ] Occurrences: 1 Examples: Occurrences: 4,342 Examples: * yüce: [j y d ʒ ɛ] * kocar: [k o d ʒ a ɾ] * edici: [e d ̪ i d ʒ ɪ] * bac: [b a d ʒ] Occurrences: 7 Examples: * hacca: [h a d ʒ ː a] |
|||||||
Sibilant |
Occurrences: 12,102 Examples: * islak: [i s ̪ ɫ a k] * asiye: [a s ̪ i j ɛ] * suyun: [s ̪ u j u n ̪] * sahir: [s ̪ a ç i ɾ] Occurrences: 115 Examples: * öss: [ø s ̪ ː] * jesse: [ʒ e s ̪ ː ɛ] * bassa: [b a s ̪ ː a] * hass: [h a s ̪ ː] Occurrences: 6,426 Examples: * azı: [a z ̪ ɨ] * bozar: [b o z ̪ a ɾ] * gezdi: [ɟ e z ̪ d ̪ ɪ] * kazı: [k a z ̪ ɨ] Occurrences: 14 Examples: * gazze: [ɡ a z ̪ ː ɛ] |
Occurrences: 6,973 Examples: * aşkın: [a ʃ k ɯ n ̪] * çitin: [t ʃ i t ̪ i n ̪] * kaşık: [k a ʃ ɯ k] * kaç: [k a t ʃ] Occurrences: 4 Examples: * eşşek: [e ʃ ː e c] Occurrences: 251 Examples: * çince: [t ʃ i n ̪ d ʒ ɛ] * casus: [d ʒ a s ̪ u s ̪] * ağaca: [a a d ʒ a] * hacmi: [h a d ʒ m ɪ] |
||||||
Fricative |
Occurrences: 1,792 Examples: * five: [f i v ɛ] * lifli: [ʎ i f ʎ ɪ] * faili: [f a i ʎ ɪ] * fakat: [f a k a t ̪] Occurrences: 25 Examples: * iffet: [i f ː e t ̪] * öff: [ø f ː] * duffy: [d ̪ u f ː i] * affet: [a f ː e t ̪] Occurrences: 2,848 Examples: * vurdu: [v u ɾ d ̪ ʊ] * yavru: [j a v ɾ ʊ] * pavel: [p a v e ʎ] * juva: [ʒ u v a] Occurrences: 44 Examples: * evvel: [e v ː e ʎ] |
Occurrences: 698 Examples: * dhkpc: [d ̪ e ç e c e p e d ʒ ɛ] * hişt: [ç i ʃ t ̪] * şahit: [ʃ a ç i t ̪] * hisle: [ç i s ̪ ʎ ɛ] Occurrences: 7 Examples: * sıhhi: [s ̪ ɯ ç ː ɪ] |
Occurrences: 2,651 Examples: * şehre: [ʃ e h ɾ ɛ] * merih: [m e ɾ i h] * tamah: [t ̪ a m a h] * vazıh: [v a z ̪ ɯ h] Occurrences: 13 Examples: * uhhuh: [u h ː u h] * öhhü: [ø h ː ʏ] * ohhü: [o h ː ʏ] * ahha: [a h ː a] |
|||||
Approximant |
Occurrences: 14,850 Examples: * ayıbı: [a j ɯ b ɨ] * ayran: [a j ɾ a n ̪] * giyer: [ɟ i j e ɾ] * şeydi: [ʃ e j d ̪ ɪ] Occurrences: 29 Examples: * ayyar: [a j ː a ɾ] * ayyaş: [a j ː a ʃ] |
Occurrences: 2,021 Examples: * boğar: [b o ɰ a ɾ] * ayağa: [a j a ɰ a] * sağım: [s ̪ a ɰ ɯ m] * yağar: [j a ɰ a ɾ] |
||||||
Tap |
Occurrences: 27,151 Examples: * prof: [p ɾ o f] * teori: [t ̪ e o ɾ ɪ] * durma: [d ̪ u ɾ m a] * sirk: [s ̪ i ɾ c] Occurrences: 51 Examples: * şerri: [ʃ e ɾ ː ɪ] * pırrr: [p ɯ ɾ ː] * şerre: [ʃ e ɾ ː ɛ] * zerre: [z ̪ e ɾ ː ɛ] |
|||||||
Lateral |
Occurrences: 14,862 Examples: * oyla: [o j ɫ a] * çaplı: [t ʃ a p ɫ ɨ] * ukala: [u k a ɫ a] * öldün: [ø ɫ d ̪ y n ̪] Occurrences: 382 Examples: * hall: [h a ɫ ː] * çallı: [t ʃ a ɫ ː ɨ] * all: [a ɫ ː] * allah: [a ɫ ː a] |
Occurrences: 10,269 Examples: * şekil: [ʃ e c i ʎ] * lere: [ʎ e ɾ ɛ] * dölü: [d ̪ ø ʎ ʏ] * kmlik: [k m ʎ i c] Occurrences: 437 Examples: * sille: [s ̪ i ʎ ː ɛ] * dilli: [d ̪ i ʎ ː ɪ] * eller: [e ʎ ː e ɾ] * kelle: [c e ʎ ː ɛ] |
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: 23,548 Examples: * inen: [i n ̪ e n ̪] * nasip: [n ̪ a s ̪ i p] * sildi: [s ̪ i ʎ i d ̪ ɪ] * fiyat: [f i j a t ̪] Occurrences: 24 Examples: * çiğli: [t ʃ i ː ʎ ɪ] * iğne: [i ː n ̪ ɛ] * niğde: [n ̪ i ː d ̪ ɛ] * çiğne: [t ʃ i ː n ̪ ɛ] Occurrences: 6,145 Examples: * ölüme: [ø ʎ y m ɛ] * düşer: [d ̪ y ʃ e ɾ] * yürü: [j y ɾ ʏ] * üye: [y j ɛ] Occurrences: 4 Examples: * düğme: [d ̪ y ː m ɛ] * büğrü: [b y ː ɾ ʏ] |
Occurrences: 5,437 Examples: * sazı: [s ̪ a z ̪ ɨ] * aşçı: [a ʃ t ʃ ɨ] * sızdı: [s ̪ ɯ z ̪ d ̪ ɨ] * başlı: [b a ʃ ɫ ɨ] |
Occurrences: 20,294 Examples: * mısın: [m ɯ s ̪ ɯ n ̪] * açın: [a t ʃ ɯ n ̪] * kırar: [k ɯ ɾ a ɾ] * yakıp: [j a k ɯ p] Occurrences: 17 Examples: * atığ: [a t ̪ ɯ ː] * çığ: [t ʃ ɯ ː] Occurrences: 11,506 Examples: * otuza: [o t ̪ u z ̪ a] * kabul: [k a b u ɫ] * şura: [ʃ u ɾ a] * tut: [t ̪ u t ̪] Occurrences: 86 Examples: * uğra: [u ː ɾ a] * huğlu: [h u ː ɫ ʊ] * tuğra: [t ̪ u ː ɾ a] * uğur: [u ː ɾ] |
||
Occurrences: 4,904 Examples: * adli: [a d ̪ ʎ ɪ] * aldi: [a ɫ d ̪ ɪ] * içini: [i t ʃ i n ̪ ɪ] * kenti: [c e n ̪ t ̪ ɪ] Occurrences: 1 Examples: * çiğ: [t ʃ ɪ ː] Occurrences: 499 Examples: * dölü: [d ̪ ø ʎ ʏ] * ürktü: [y ɾ c t ̪ ʏ] * görkü: [ɟ ø ɾ c ʏ] * üzdü: [y z ̪ d ̪ ʏ] |
Occurrences: 1,502 Examples: * doru: [d ̪ o ɾ ʊ] * huğlu: [h u ː ɫ ʊ] * kodu: [k o d ̪ ʊ] * doğru: [d ̪ o ː ɾ ʊ] |
||||
Close-Mid |
Occurrences: 25,601 Examples: * tesav: [t ̪ e s ̪ a v] * türev: [t ̪ y ɾ e v] * yense: [j e n ̪ s ̪ ɛ] * merak: [m e ɾ a k] Occurrences: 69 Examples: * eğri: [e ː ɾ ɪ] * değme: [d ̪ e ː m ɛ] * eğdi: [e ː d ̪ ɪ] * değse: [d ̪ e ː s ̪ ɛ] Occurrences: 2,725 Examples: * öküz: [ø c y z ̪] * öbek: [ø b e c] * dönüm: [d ̪ ø n ̪ y m] * görme: [ɟ ø ɾ m ɛ] Occurrences: 113 Examples: * öğle: [ø ː ʎ ɛ] * öğlen: [ø ː ʎ e n ̪] * öğren: [ø ː ɾ e n ̪] |
Occurrences: 8,835 Examples: * hoş: [h o ʃ] * ikona: [i k o n ̪ a] * toklu: [t ̪ o k ɫ ʊ] * credo: [d ʒ ɾ e d ̪ o] Occurrences: 153 Examples: * oğlum: [o ː ɫ u m] * oğlak: [o ː ɫ a k] * oğlan: [o ː ɫ a n ̪] * doğru: [d ̪ o ː ɾ ʊ] |
|||
Open-Mid |
Occurrences: 6,832 Examples: * müge: [m y ɟ ɛ] * gece: [ɟ e d ʒ ɛ] * kelt: [c ɛ ʎ t ̪] * sende: [s ̪ e n ̪ d ̪ ɛ] Occurrences: 4 Examples: * fetö: [f e t ̪ œ] * fkö: [f e c e œ] * dtö: [d ̪ e t ̪ e œ] |
||||
Open |
Occurrences: 47,570 Examples: * şan: [ʃ a n ̪] * mazur: [m a z ̪ u ɾ] * kaz: [k a z ̪] * lavaş: [ɫ a v a ʃ] Occurrences: 407 Examples: * sağcı: [s ̪ a ː d ʒ ɨ] * bağ: [b a ː] * saat: [s ̪ a ː t ̪] * ağza: [a ː z ̪ a] |