Czech MFA dictionary v2.0.0a#
@techreport{mfa_czech_mfa_dictionary_2022,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Czech MFA dictionary v2.0.0a},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Czech/Czech 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 czech_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 Czech transcripts.
This dictionary uses the MFA phone set for Czech, and was used in training the Czech 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 |
Alveolar |
Alveopalatal |
Palatal |
Velar |
Glottal |
---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 28,229 Examples: * mění: [m ɲ ɛ ɲ i ː] * rámci: [r a ː m t s ɪ] * mylné: [m ɪ l n ɛ ː] * řádem: [r ̝ a ː d ɛ m] Occurrences: 514 Examples: * změn: [z m ̩ ɲ ɛ n] * směně: [s m ̩ ɲ ɛ ɲ ɛ] * smědý: [s m ̩ ɲ ɛ d i ː] * směju: [s m ̩ ɲ ɛ j u] |
Occurrences: 376 Examples: * trumf: [t r u ɱ f] * enf: [ɛ ɱ f] * enver: [ɛ ɱ v ɛ r] * pánve: [p a ː ɱ v ɛ] |
Occurrences: 43,914 Examples: * allen: [a l ɛ n] * foton: [f o t o n] * topná: [t o p n a ː] * tajný: [t a j n i ː] Occurrences: 28 Examples: * srnku: [s r n ̩ k u] * trnka: [t r n ̩ k a] * hrnci: [ɦ r n ̩ t s ɪ] * srnka: [s r n ̩ k a] |
Occurrences: 17,203 Examples: * stráň: [s t r a ː ɲ] * páně: [p a ː ɲ ɛ] * číně: [t ʃ i ː ɲ ɛ] * měním: [m ɲ ɛ ɲ i ː m] |
Occurrences: 1,684 Examples: * hanku: [ɦ a ŋ k u] * wong: [v o ŋ k] * nkú: [ŋ k u ː] * wonky: [v o ŋ k ɪ] |
||
Stop |
Occurrences: 29,124 Examples: * hikep: [ɦ ɪ k ɛ p] * spisy: [s p ɪ s ɪ] * pluk: [p l u k] * pášu: [p a ː ʃ u] Occurrences: 11,337 Examples: * hrobu: [ɦ r o b u] * vazbu: [v a z b u] * sbd: [ɛ s b ɛ ː d ɛ ː] * šibal: [ʃ ɪ b a l] |
Occurrences: 41,150 Examples: * potká: [p o t k a ː] * žít: [ʒ i ː t] * zvíci: [z v i ː t s ɪ] * trupe: [t r u p ɛ] Occurrences: 20,430 Examples: * vláda: [v l a ː d a] * dubá: [d u b a ː] * dipól: [d ɪ p o ː l] * jedva: [j ɛ d v a] |
Occurrences: 5,791 Examples: * pěsti: [p j ɛ s c ɪ] * zátěž: [z a ː c ɛ ʃ] * sleť: [s l ɛ c] * loď: [l o c] Occurrences: 3,394 Examples: * radíc: [r a ɟ i ː t s] * lodě: [l o ɟ ɛ] * láďa: [l a ː ɟ a] * řadíc: [r ̝ a ɟ i ː t s] |
Occurrences: 33,795 Examples: * otko: [o t k o] * úzkým: [u ː s k i ː m] * kec: [k ɛ t s] * kolky: [k o l k ɪ] Occurrences: 4,303 Examples: * gero: [ɡ ɛ r o] * golfu: [ɡ o l f u] * droga: [d r o ɡ a] * švagr: [ʃ v a ɡ r ̩] |
|||
Affricate |
Occurrences: 14,121 Examples: * děcka: [ɟ ɛ t s k a] * čtecí: [t ʃ t ɛ t s i ː] * celku: [t s ɛ l k u] * lance: [l a n t s ɛ] |
Occurrences: 8,670 Examples: * tlačí: [t l a t ʃ i ː] * mačka: [m a t ʃ k a] * čeští: [t ʃ ɛ ʃ c i ː] * čadit: [t ʃ a ɟ ɪ t] Occurrences: 148 Examples: * fidži: [f ɪ d ʒ ɪ] * juda: [d ʒ u d a] * džob: [d ʒ o p] * padžo: [p a d ʒ o] |
|||||
Sibilant |
Occurrences: 34,653 Examples: * texas: [t ɛ k s a s] * servu: [s ɛ r v u] * ctěn: [t s c ɛ n] * slize: [s l ɪ z ɛ] Occurrences: 16,421 Examples: * krizi: [k r ɪ z ɪ] * brzdí: [b r ̩ z ɟ i ː] * eznam: [ɛ z n a m] * zdrhl: [z d r ̩ ɦ l ̩] |
Occurrences: 7,467 Examples: * šídlo: [ʃ i ː d l o] * lukáš: [l u k a ː ʃ] * čočky: [t ʃ o t ʃ k ɪ] * aše: [a ʃ ɛ] Occurrences: 4,716 Examples: * nížit: [ɲ i ː ʒ ɪ t] * ržát: [r ʒ a ː t] * řežu: [r ̝ ɛ ʒ u] * žabí: [ʒ a b i ː] |
|||||
Fricative |
Occurrences: 5,655 Examples: * filc: [f ɪ l t s] * savče: [s a f t ʃ ɛ] * mafře: [m a f r ̝ ɛ] * bow: [b o f] Occurrences: 36,539 Examples: * úvahy: [u ː v a ɦ ɪ] * vršek: [v r ̩ ʃ ɛ k] * výdaj: [v i ː d a j] * hovoř: [ɦ o v o r ̝] |
Occurrences: 11,488 Examples: * hladě: [ɦ l a ɟ ɛ] * hrát: [ɦ r a ː t] * hladu: [ɦ l a d u] * hlíny: [ɦ l i ː n ɪ] |
|||||
Approximant |
Occurrences: 16,040 Examples: * juana: [j u a n a] * libyi: [l ɪ b a j ɪ] * youth: [j o w t] * vědro: [v j ɛ d r o] |
||||||
Trill |
Occurrences: 33,727 Examples: * fort: [f o r t] * vran: [v r a n] * dráze: [d r a ː z ɛ] * vření: [v r ̝ ɛ ɲ i ː] Occurrences: 2,240 Examples: * fofr: [f o f r ̩] * prče: [p r ̩ t ʃ ɛ] * mutr: [m u t r ̩] * mrzly: [m r ̩ z l ɪ] |
||||||
Lateral |
Occurrences: 38,616 Examples: * lví: [l v i ː] * popel: [p o p ɛ l] * liják: [l ɪ j a ː k] * ujala: [u j a l a] Occurrences: 977 Examples: * jedl: [j ɛ d l ̩] * mohl: [m o ɦ l ̩] * núl: [n u ː l ̩] * blbý: [b l ̩ b i ː] |
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: 0 Examples: * klíčů: [k l i ː t ʃ u ː] * luční: [l u t ʃ ɲ i ː] * bdím: [b ɟ i ː m] * sazím: [s a z i ː m] Occurrences: 35,583 Examples: * ubozí: [u b o z i ː] * tkání: [t k a ː ɲ i ː] * nouzí: [n o w z i ː] * alího: [a l i ː ɦ o] |
Occurrences: 21,350 Examples: * útlou: [u ː t l o w] * ujeď: [u j ɛ c] * pudl: [p u d l ̩] * lídrů: [l i ː d r u ː] Occurrences: 5,296 Examples: * kosům: [k o s u ː m] * úvaha: [u ː v a ɦ a] * honů: [ɦ o n u ː] * rubům: [r u b u ː m] |
|||
Occurrences: 56,634 Examples: * taxy: [t a k s ɪ] * hrůzy: [ɦ r u ː z ɪ] * vytí: [v ɪ c i ː] * burzy: [b u r z ɪ] |
|||||
Close-Mid |
Occurrences: 64,660 Examples: * slow: [s l o f] * poldi: [p o l ɟ ɪ] * koťat: [k o c a t] * hole: [ɦ o l ɛ] Occurrences: 6,094 Examples: * four: [f o w r] * rudou: [r u d o w] * pekou: [p ɛ k o w] * cloud: [k l o w t] Occurrences: 399 Examples: * tón: [t o ː n] * glóbu: [ɡ l o ː b u] * glóbe: [ɡ l o ː b ɛ] * jógou: [j o ː ɡ o w] |
||||
Open-Mid |
Occurrences: 70,232 Examples: * nebýt: [n ɛ b i ː t] * slepě: [s l ɛ p j ɛ] * nežli: [n ɛ ʒ l ɪ] * vázne: [v a ː z n ɛ] Occurrences: 9,277 Examples: * dph: [d ɛ ː p ɛ ː ɦ a ː] * orné: [o r n ɛ ː] * renée: [r ɛ n ɛ ː] * bové: [b o v ɛ ː] |
||||
Open |
Occurrences: 57,350 Examples: * diva: [d ɪ v a] * ďáblů: [ɟ a ː b l u ː] * čapím: [t ʃ a p i ː m] * újma: [u ː j m a] Occurrences: 20,945 Examples: * skály: [s k a ː l ɪ] * žábry: [ʒ a ː b r ɪ] * nedbá: [n ɛ d b a ː] * urát: [u r a ː t] |
Diphthongs#
aw
ew
ow