Turkish MFA dictionary v2.0.0#
@techreport{mfa_turkish_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Turkish MFA dictionary v2.0.0},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Turkish/Turkish MFA dictionary v2_0_0.html}},
year={2022},
month={Mar},
}
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: 17,544 Examples: * derim: [d̪ e ɾ i m] * demir: [d̪ e m i ɾ] * ölmüş: [ø ʎ m y ʃ] * aklım: [a k ɫ ɯ m] Occurrences: 51 Examples: * hımm: [h ɯ mː] * amman: [a mː a n̪] * emme: [e mː ɛ] * emmiş: [e mː i ʃ] |
Occurrences: 26,972 Examples: * konan: [k o n̪ a n̪] * ajans: [a ʒ a n̪ s̪] * nişan: [n̪ i ʃ a n̪] * seven: [s̪ e v e n̪] Occurrences: 76 Examples: * annem: [a n̪ː e m] * ninni: [n̪ i n̪ː ɪ] * annen: [a n̪ː e n̪] * anne: [a n̪ː ɛ] |
Occurrences: 189 Examples: * engin: [ɛ ɲ ɟ i n̪] * günkü: [ɟ y ɲ c ʏ] * denge: [d̪ ɛ ɲ ɟ ɛ] * dünkü: [d̪ y ɲ c ʏ] |
Occurrences: 205 Examples: * yonga: [j o ŋ ɡ a] * ongun: [o ŋ ɡ u n̪] * yankı: [j a ŋ k ɨ] * tank: [t̪ a ŋ k] |
||||
Stop |
Occurrences: 3,311 Examples: * bbp: [b e b e p ɛ] * dikip: [d̪ i c i p] * piç: [p i tʃ] * pelin: [p e ʎ i n̪] Occurrences: 5 Examples: * zappa: [z̪ a pː a] * züppe: [z̪ y pː ɛ] * kappa: [k a pː a] Occurrences: 6,146 Examples: * ibrik: [i b ɾ i k] * bilgi: [b i ʎ ɟ ɪ] * abdli: [a b d̪ ʎ ɪ] * berat: [b e ɾ a t̪] Occurrences: 22 Examples: * kubbe: [k u bː ɛ] * abbas: [a bː a s̪] * tıbbi: [t̪ ɯ bː ɪ] |
Occurrences: 13,331 Examples: * berat: [b e ɾ a t̪] * vatan: [v a t̪ a n̪] * tuzla: [t̪ u z̪ ɫ a] * tirit: [t̪ i ɾ i t̪] Occurrences: 629 Examples: * attı: [a t̪ː ɨ] * ettin: [e t̪ː i n̪] * örttü: [ø ɾ t̪ː ʏ] * çattı: [tʃ a t̪ː ɨ] Occurrences: 16,587 Examples: * derim: [d̪ e ɾ i m] * demir: [d̪ e m i ɾ] * dik: [d̪ i c] * abdli: [a b d̪ ʎ ɪ] Occurrences: 139 Examples: * abdde: [a b d̪ː ɛ] * redde: [ɾ e d̪ː ɛ] * maddi: [m a d̪ː ɪ] * addır: [a d̪ː ɯ ɾ] |
Occurrences: 5,935 Examples: * dik: [d̪ i c] * ürkek: [y ɾ c e c] * şevk: [ʃ e v c] * çörek: [tʃ ø ɾ e c] Occurrences: 24 Examples: * sikke: [s̪ i cː ɛ] * ökkeş: [ø cː e ʃ] * takke: [t̪ a cː ɛ] Occurrences: 3,319 Examples: * bilgi: [b i ʎ ɟ ɪ] * girip: [ɟ i ɾ i p] * güneş: [ɟ y n̪ e ʃ] * güya: [ɟ y j a] |
Occurrences: 11,219 Examples: * ibrik: [i b ɾ i k] * konan: [k o n̪ a n̪] * aklım: [a k ɫ ɯ m] * uçak: [u tʃ a k] Occurrences: 84 Examples: * dakka: [d̪ a kː a] * akka: [a kː a] * hakkı: [h a kː ɨ] * okka: [o kː a] Occurrences: 678 Examples: * yonga: [j o ŋ ɡ a] * gamlı: [ɡ a m ɫ ɨ] * saygı: [s̪ a j ɡ ɨ] * algı: [a ɫ ɡ ɨ] Occurrences: 2 Examples: |
||||
Affricate |
Occurrences: 3,644 Examples: * uçak: [u tʃ a k] * çaba: [tʃ a b a] * kulaç: [k u ɫ a tʃ] * çörek: [tʃ ø ɾ e c] Occurrences: 1 Examples: Occurrences: 4,336 Examples: * cihan: [dʒ i h a n̪] * ceviz: [dʒ e v i z̪] * cnn: [dʒ e n̪ e n̪ ɛ] * clan: [dʒ ɫ a n̪] Occurrences: 7 Examples: * hacca: [h a dʒː a] |
|||||||
Sibilant |
Occurrences: 12,071 Examples: * sığar: [s̪ ɯ a ɾ] * asma: [a s̪ m a] * ajans: [a ʒ a n̪ s̪] * seven: [s̪ e v e n̪] Occurrences: 115 Examples: * üssü: [y s̪ː e s̪ e ʏ] * hissi: [ç i s̪ː ɪ] * üsse: [y s̪ː ɛ] * ıssız: [ɯ s̪ː ɯ z̪] Occurrences: 6,418 Examples: * arz: [a ɾ z̪] * tuzla: [t̪ u z̪ ɫ a] * hızlı: [h ɯ z̪ ɫ ɨ] * ceviz: [dʒ e v i z̪] Occurrences: 14 Examples: * gazze: [ɡ a z̪ː ɛ] |
Occurrences: 6,959 Examples: * ölmüş: [ø ʎ m y ʃ] * nişan: [n̪ i ʃ a n̪] * şevk: [ʃ e v c] * işlek: [i ʃ ʎ e k] Occurrences: 4 Examples: * eşşek: [e ʃː e c] Occurrences: 251 Examples: * ajans: [a ʒ a n̪ s̪] * tiraj: [t̪ i ɾ a ʒ] * lajiç: [ɫ a ʒ i tʃ] * jokiç: [ʒ o k tʃ] |
||||||
Fricative |
Occurrences: 1,789 Examples: * arefe: [a ɾ e f ɛ] * kefil: [c e f i ʎ] * zarf: [z̪ a ɾ f] * phen: [f e n̪] Occurrences: 25 Examples: * affet: [a fː e t̪] * duffy: [d̪ u fː i] * öff: [ø fː] * iffet: [i fː e t̪] Occurrences: 2,841 Examples: * seven: [s̪ e v e n̪] * vatan: [v a t̪ a n̪] * avni: [a v n̪ ɪ] * şevk: [ʃ e v c] Occurrences: 44 Examples: * evvel: [e vː e ʎ] |
Occurrences: 692 Examples: * yohei: [j o ç e ɪ] * şüphe: [ʃ y p ç ɛ] * chp: [dʒ e ç e p ɛ] * hekim: [ç e c i m] Occurrences: 7 Examples: * sıhhi: [s̪ ɯ çː ɪ] |
Occurrences: 2,646 Examples: * havlu: [h a v ɫ ʊ] * cihan: [dʒ i h a n̪] * hızlı: [h ɯ z̪ ɫ ɨ] * sulh: [s̪ u ɫ h] Occurrences: 13 Examples: * ahhh: [a hː] * ahha: [a hː a] * ahh: [a hː] * ohhü: [o hː ʏ] |
|||||
Approximant |
Occurrences: 14,809 Examples: * yonga: [j o ŋ ɡ a] * eyi: [e j ɪ] * ebeyi: [e b e j ɪ] * sayan: [s̪ a j a n̪] Occurrences: 29 Examples: * ayyaş: [a jː a ʃ] * ayyar: [a jː a ɾ] |
Occurrences: 2,020 Examples: * sığar: [s̪ ɯ ɰ a ɾ] * sağıp: [s̪ a ɰ ɯ p] * aşağı: [a ʃ a ɰ ɨ] * dağı: [d̪ a ɰ ɨ] |
||||||
Tap |
Occurrences: 27,092 Examples: * sığar: [s̪ ɯ a ɾ] * ibrik: [i b ɾ i k] * derim: [d̪ e ɾ i m] * demir: [d̪ e m i ɾ] Occurrences: 50 Examples: * harry: [h a ɾː i] * sırrı: [s̪ ɯ ɾː ɨ] * şerre: [ʃ e ɾː ɛ] * şerri: [ʃ e ɾː ɪ] |
|||||||
Lateral |
Occurrences: 14,836 Examples: * aklım: [a k ɫ ɯ m] * muğla: [m uː ɫ a] * tuzla: [t̪ u z̪ ɫ a] * kulaç: [k u ɫ a tʃ] Occurrences: 379 Examples: * kolla: [k o ɫː a] * all: [a ɫː] * pullu: [p u ɫː ʊ] * çullu: [tʃ u ɫː ʊ] |
Occurrences: 10,255 Examples: * ölmüş: [ø ʎ m y ʃ] * bilgi: [b i ʎ ɟ ɪ] * abdli: [a b d̪ ʎ ɪ] * pelin: [p e ʎ i n̪] Occurrences: 436 Examples: * kelle: [c e ʎː ɛ] * belli: [b e ʎː ɪ] * milli: [m i ʎː ɪ] * bill: [b i ʎː] |
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: 23,514 Examples: * ibrik: [i b ɾ i k] * derim: [d̪ e ɾ i m] * demir: [d̪ e m i ɾ] * dik: [d̪ i c] Occurrences: 24 Examples: * iğne: [iː n̪ ɛ] * iğnem: [iː n̪ e m] * çiğli: [tʃ iː ʎ ɪ] * çiğne: [tʃ iː n̪ ɛ] Occurrences: 6,140 Examples: * ölmüş: [ø ʎ m y ʃ] * ürkek: [y ɾ c e c] * kütle: [c y t̪ ʎ ɛ] * tüh: [t̪ y h] Occurrences: 4 Examples: * düğme: [d̪ yː m ɛ] * büğrü: [b yː ɾ ʏ] |
Occurrences: 5,424 Examples: * attı: [a t̪ː ɨ] * hızlı: [h ɯ z̪ ɫ ɨ] * aşağı: [a ʃ a ɨ] * kızı: [k ɯ z̪ ɨ] |
Occurrences: 20,253 Examples: * sığar: [s̪ ɯ a ɾ] * aklım: [a k ɫ ɯ m] * hızlı: [h ɯ z̪ ɫ ɨ] * aldık: [a ɫ d̪ ɯ k] Occurrences: 17 Examples: * atığ: [a t̪ ɯː] * çığ: [tʃ ɯː] Occurrences: 11,488 Examples: * uçak: [u tʃ a k] * tuzla: [t̪ u z̪ ɫ a] * kulaç: [k u ɫ a tʃ] * sulh: [s̪ u ɫ h] Occurrences: 85 Examples: * muğla: [m uː ɫ a] * uğur: [uː ɾ] * tuğra: [t̪ uː ɾ a] * uğru: [uː ɾ ʊ] |
||
Occurrences: 4,896 Examples: * bilgi: [b i ʎ ɟ ɪ] * abdli: [a b d̪ ʎ ɪ] * derdi: [d̪ ɛ ɾ d̪ ɪ] * eyi: [e j ɪ] Occurrences: 1 Examples: * çiğ: [tʃ ɪː] Occurrences: 499 Examples: * gözü: [ɟ ø z̪ ʏ] * sürgü: [s̪ y ɾ ɟ ʏ] * üssü: [y s̪ː e s̪ e ʏ] * türkü: [t̪ y ɾ c ʏ] |
Occurrences: 1,502 Examples: * havlu: [h a v ɫ ʊ] * bursu: [b u ɾ s̪ ʊ] * tuzlu: [t̪ u z̪ ɫ ʊ] * kumlu: [k u m ɫ ʊ] |
||||
Close-Mid |
Occurrences: 25,525 Examples: * derim: [d̪ e ɾ i m] * demir: [d̪ e m i ɾ] * arefe: [a ɾ e f ɛ] * berat: [b e ɾ a t̪] Occurrences: 69 Examples: * değme: [d̪ eː m ɛ] * değdi: [d̪ eː d̪ ɪ] * eğri: [eː ɾ ɪ] * eğdi: [eː d̪ ɪ] Occurrences: 2,718 Examples: * ölmüş: [ø ʎ m y ʃ] * çörek: [tʃ ø ɾ e c] * öib: [ø b] * öveç: [ø v e tʃ] Occurrences: 113 Examples: * öğle: [øː ʎ ɛ] * öğlen: [øː ʎ e n̪] * öğren: [øː ɾ e n̪] |
Occurrences: 8,817 Examples: * yonga: [j o ŋ ɡ a] * konan: [k o n̪ a n̪] * yohei: [j o ç e ɪ] * koy: [k o j] Occurrences: 153 Examples: * oğlak: [oː ɫ a k] * oğlan: [oː ɫ a n̪] * doğru: [d̪ oː ɾ ʊ] * doğsa: [d̪ oː s̪ a] |
|||
Open-Mid |
Occurrences: 6,815 Examples: * arefe: [a ɾ e f ɛ] * derdi: [d̪ ɛ ɾ d̪ ɪ] * bbp: [b e b e p ɛ] * bende: [b e n̪ d̪ ɛ] Occurrences: 4 Examples: * dtö: [d̪ e t̪ e œ] * fetö: [f e t̪ œ] * fkö: [f e c e œ] |
||||
Open |
Occurrences: 47,475 Examples: * sığar: [s̪ ɯ a ɾ] * yonga: [j o ŋ ɡ a] * attı: [a t̪ː ɨ] * konan: [k o n̪ a n̪] Occurrences: 407 Examples: * dağda: [d̪ aː d̪ a] * bağla: [b aː ɫ a] * ağla: [aː ɫ a] * yağlı: [j aː ɫ ɨ] |