Swedish MFA dictionary v3.0.0#
@techreport{mfa_swedish_mfa_dictionary_2024,
author={McAuliffe, Michael and Sonderegger, Morgan},
title={Swedish MFA dictionary v3.0.0},
address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Swedish/Swedish MFA dictionary v3_0_0.html}},
year={2024},
month={Mar},
}
G2P models Acoustic models |
Installation#
Install from the MFA command line:
mfa model download dictionary swedish_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](https://raw.githubusercontent.com/MontrealCorpusTools/mfa-models/main/dictionary/swedish/mfa/Swedish MFA dictionary v3_0_0.dict).
Intended use#
This dictionary is intended for forced alignment of Swedish transcripts.
This dictionary uses the MFA phone set for Swedish, and was used in training the Swedish 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 |
Retroflex |
Palatal |
Velar |
Glottal |
---|---|---|---|---|---|---|---|---|
Nasal |
Occurrences: 11,523 Examples: * magin: [m a ɡ ɪ n̪] * töm: [t̪ʰ øː m] * mask: [m a s̪ k] * temat: [t̪ʰ eː m ɑː t̪] |
Occurrences: 32,592 Examples: * renas: [r eː˧˩ n̪ a˥˩ s̪] * fin: [f iː n̪] * latin: [l̪ ɑː t̪ ɪ n̪] * sänt: [s̪ ɛ n̪ t̪] |
Occurrences: 1,916 Examples: * öarna: [øː a ɳ a] * åarna: [ɔ˧˩ a˥˩ ɳ a] * garn: [ɡ a ɳ] * torn: [t̪ʰ uː ɳ] |
Occurrences: 6,056 Examples: * lunga: [l̪ ɵ˧˩ ŋ a˥˩] * flink: [f l̪ ɪ ŋ k] * klang: [k l̪ a ŋ] * tung: [t̪ʰ ɵ ŋ] |
||||
Stop Plain |
Occurrences: 5,885 Examples: * prydd: [p r ʏ d̪] * pröva: [p r øː ʋ a] * uppnå: [ɵ˧˩ p n̪ oː] * söp: [s̪ øː p] Occurrences: 7,819 Examples: * fbi: [ɛ f b eː iː] * breda: [b r eː d̪ a] * bakas: [b ɑː k a s̪] * bön: [b øː n̪] |
Occurrences: 26,381 Examples: * bildt: [b ɪ l̪ d̪ t̪] * vuxit: [ʋ ɵ k s̪ ɪ t̪] * act: [a k t̪] * benet: [b ɛ n̪ ɛ t̪] Occurrences: 16,432 Examples: * ålagd: [oː˧˩ l̪ a˥˩ j d̪] * utbud: [ʉː˧˩ t̪ b ʉː d̪] * oreda: [ʊ r eː d̪ a] * ryd: [r yː d̪] |
Occurrences: 1,084 Examples: * rart: [r ɑː ʈ] * årtal: [oː ʈ ɑː l̪] * fort: [f ʊ ʈ] * kvart: [k ʋ ɑː ʈ] Occurrences: 1,428 Examples: * förde: [f œ ɖ ɛ] * börda: [b øː ɖ a] * flärd: [f l̪ ɛː ɖ] * order: [ɔ ɖ ɛ r] |
Occurrences: 14,896 Examples: * äkta: [ɛ k t̪ a] * drick: [d̪ r ɪ k] * klapp: [k l̪ a˥˩ p] * etisk: [eː t̪ ɪ s̪ k] Occurrences: 8,422 Examples: * liga: [l̪ ɪ ɡ a] * gorgo: [ɡ ʊ r ɡ ɔ] * kaxig: [kʰ a k s̪ ɪ ɡ] * grävs: [ɡ r ɛː ʋ s̪] |
||||
Aspirated |
Occurrences: 2,401 Examples: * pekas: [pʰ eː k a˥˩ s̪] * kopia: [kʰ ʊ pʰ iː˧˩ a˥˩] * pizza: [pʰ ɪ˧˩ t̪ s̪ a˥˩] * pma: [pʰ eː ɛ m ɑː] |
Occurrences: 5,902 Examples: * råtta: [r ɔ t̪ʰ a] * betet: [b ɛ t̪ʰ eː t̪] * tona: [t̪ʰ uː˧˩ n̪ a˥˩] * tag: [t̪ʰ ɑː ɡ] |
Occurrences: 321 Examples: * orten: [ʊ ʈʰ eː n̪] * ärter: [ɛ ʈʰ eː r] * orter: [ʊ ʈʰ eː r] * märta: [m ɛ ʈʰ a] |
Occurrences: 3,438 Examples: * kosan: [kʰ ɔ s̪ a n̪] * kol: [kʰ oː l̪] * kult: [kʰ ɵ l̪ t̪] * hacka: [h a kʰ a] |
||||
Sibilant |
Occurrences: 30,931 Examples: * bloss: [b l̪ ɔ s̪] * feins: [f a j n̪ s̪] * skalv: [s̪ k a˥˩ l̪ ʋ] * söta: [s̪ øː t̪ a˥˩] |
Occurrences: 1,872 Examples: * bärs: [b ɛː ʂ] * skärs: [ɧ ɛː ʂ] * pärs: [pʰ ɛː ʂ] * vars: [ʋ a ʂ] |
Occurrences: 1,044 Examples: * köpas: [ɕ øː pʰ a s̪] * köp: [ɕ øː p] * köken: [ɕ øː k ɛ n̪] * köpen: [ɕ øː p ɛ n̪] |
|||||
Fricative |
Occurrences: 9,152 Examples: * först: [f œ ʂ t̪] * femte: [f ɛ n̪ t̪ʰ ɛ] * filma: [f ɪ l̪ m a] * flög: [f l̪ øː ɡ] |
Occurrences: 5,278 Examples: * huset: [h ʉː s̪ ɛ t̪] * hyran: [h ʏ r ɑː n̪] * hel: [h eː l̪] * happ: [h a p] |
||||||
Approximant |
Occurrences: 9,674 Examples: * kval: [k ʋ ɑː l̪] * yngve: [ʏ˧˩ ŋ ʋ ɛ] * kurva: [kʰ ɵ˥˩ r ʋ a] * snuva: [s̪ n̪ ʉː ʋ a] |
Occurrences: 4,540 Examples: * organ: [ɔ r j a˥˩ n̪] * bjärt: [b j ɛ ʈ] * jakt: [j a k t̪] * hej: [h ɛ j] |
||||||
Trill |
Occurrences: 36,046 Examples: * boris: [b uː r iː s̪] * tåras: [t̪ʰ oː r a s̪] * betar: [b eː t̪ a r] * dator: [d̪ ɑː˧˩ t̪ ɔ˥˩ r] |
|||||||
Lateral |
Occurrences: 21,162 Examples: * slam: [s̪ l̪ ɑː m] * lilla: [l̪ ɪ l̪ a] * orgel: [ɔ r j ɛ l̪] * flöt: [f l̪ øː t̪] |
|||||||
Lateral Tap |
Occurrences: 353 Examples: * gurli: [ɡ ɵ ɭ ɪ] * jarl: [j ɑː ɭ] * jarro: [j ɑː ɭ s̪ r ʊ] * ärla: [ɛː ɭ a] |
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: Occurrences: 5,715 Examples: * idre: [iː˧˩ d̪ r ɛ˥˩] * nivå: [n̪ iː ʋ oː] * visor: [ʋ iː˥˩ s̪ uː r] * vidga: [ʋ iː˧˩ d̪ ɡ a] Occurrences: 0 Examples: Occurrences: 1,431 Examples: * otyg: [uː˧˩ t̪ʰ yː ɡ] * psyke: [p s̪ yː˥˩ k ɛ˥˩] * fyra: [f yː r a] * bryt: [b r yː˧˩ t̪] |
Occurrences: 0 Examples: Occurrences: 3,567 Examples: * hukar: [h ʉː˥˩ k a r] * utöva: [ʉː t̪ øː ʋ a] * bruka: [b r ʉː˥˩ k a] * ssu: [ɛ s̪ ɛ s̪ ʉː] |
Occurrences: 0 Examples: Occurrences: 4,520 Examples: * slogs: [s̪ l̪ uː ɡ s̪] * floyd: [f l̪ uː yː d̪] * lågor: [l̪ oː ɡ uː r] * spola: [s̪ p uː˥˩ l̪ a] |
||
Occurrences: 19,020 Examples: * misär: [m ɪ˧˩ s̪ ɛː r] * ibsen: [ɪ˧˩ b s̪ ɛ n̪] * exil: [ɛ k s̪ ɪ˩ l̪] * tokig: [t̪ʰ uː k ɪ˥˩ ɡ] Occurrences: 2,485 Examples: * kyl: [ɕ ʏ l̪] * tyskt: [t̪ʰ ʏ s̪ t̪] * donny: [d̪ ɔ n̪ ʏ˩] * ronny: [r ɔ n̪ ʏ˩] |
Occurrences: 5,336 Examples: * ohyra: [ʊ h yː r a] * älgko: [ɛ l̪ j k ʊ˩] * dator: [d̪ ɑː˧˩ t̪ ʊ˥˩ r] * judo: [j ʉː˧˩ d̪ ʊ] |
||||
Close-Mid |
Occurrences: 0 Examples: Occurrences: 9,134 Examples: * rpm: [ɛ r pʰ eː ɛ m] * vinet: [ʋ ɪ n̪ eː t̪] * leva: [l̪ eː˧˩ ʋ a˥˩] * aceas: [a k eː˧˩ a˥˩ s̪] Occurrences: 0 Examples: Occurrences: 4,406 Examples: * bröms: [b r øː˥˩ m s̪] * sökas: [s̪ øː˧˩ k a˥˩ s̪] * bössa: [b øː˥˧ ɑː] * ornö: [ʊ r n̪ øː] |
Occurrences: 7,028 Examples: * tungt: [t̪ʰ ɵ˧˩ n̪ t̪] * wulf: [ʋ ɵ l̪ f] * musik: [m ɵ˥˩ s̪ iː k] * tunga: [t̪ʰ ɵ˧˩ ŋ a˥˩] |
Occurrences: 0 Examples: Occurrences: 4,039 Examples: * åar: [oː˧˩ ɑː r] * inpå: [ɪ˧˩ m p oː˥˩] * åtala: [oː˧˩ t̪ a˥˩ l̪ a] * fåror: [f oː˧˩ r uː r] |
||
Open-Mid |
Occurrences: 39,571 Examples: * mäns: [m ɛ˧˩ n̪ s̪] * häpen: [h ɛ p ɛ˩ n̪] * svepa: [s̪ ʋ ɛ˥˧ p a] * fräs: [f r ɛ˧˩ s̪] Occurrences: 2,788 Examples: * aero: [a ɛː r ʊ] * fläta: [f l̪ ɛː˥˩ t̪ a˥˩] * iväg: [iː ʋ ɛː˧˩ ɡ] * bärga: [b ɛː˥˩ r ɡ a] Occurrences: 2,797 Examples: * vinön: [ʋ ɪ˧˩ n̪ œ n̪] * möbel: [m œ˧˩ b ɛ l̪] * stört: [s̪ t̪ œ˥˩ ʈ] * öppet: [œ˧˩ p ɛ˥˩ t̪] |
Occurrences: 8,352 Examples: * osedd: [ɔ˥˧ s̪ eː d̪] * elof: [ɛ l̪ ɔ f] * rock: [r ɔ˥˧ k] * world: [ʋ ɔ r l̪ d̪] |
|||
Open |
Occurrences: 34,033 Examples: * samla: [s̪ a˧˩ m l̪ a] * max: [m a˥˧ k s̪] * greta: [ɡ r eː˧˩ t̪ a˥˩] * albin: [a˥˧ l̪ b ɪ˩ n̪] |
Occurrences: 0 Examples: Occurrences: 10,326 Examples: * fåtal: [f oː˧˩ t̪ ɑː˥˩ l̪] * tage: [t̪ʰ ɑː ɧ] * dödat: [d̪ øː˧˩ d̪ ɑː t̪] * avge: [ɑː˧˩ ʋ ɡ ɛ˥˩] |
Tones#
˥˧
˥˩
˧˩
˩