Ctrl+K
Ctrl+K

Kurmanji CV dictionary v2.0.0#

  • Maintainer: Vox Communis

  • Language: Kurmanji

  • Dialect: N/A

  • Phone set: Epitran

  • Number of words: 10,771

  • Phones: aː b d d͡ʒ eː f h iː j k l m n o p q s t t͡ʃ u uː v w x z ŋ ɛ ɡ ɣ ɪ ɫ ɾ ʃ ʏ ʒ

  • License: CC-0

  • Compatible MFA version: v2.0.0

  • Citation:

@misc{Ahn_Chodroff_2022,
	author={Ahn, Emily and Chodroff, Eleanor},
	title={VoxCommunis Corpus},
	address={\url{https://osf.io/t957v}},
	publisher={OSF},
	year={2022},
	month={Jan}
}

Acoustic models

Installation#

Install from the MFA command line:

mfa model download dictionary kurmanji_cv

Or download from the release page.

Intended use#

This dictionary is intended for forced alignment of Kurmanji transcripts.

This dictionary uses the Epitran phone set for Kurmanji, and was used in training the Kurmanji Epitran 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

Uvular

Glottal

Nasal
Occurrences:
1,921
Examples:
* "em:
[ɛ m]
* mela:
[m ɛ l aː]
* çeman:
[t͡ʃ ɛ m aː n]
* maç,:
[m aː t͡ʃ]
Occurrences:
6,033
Examples:
* pênc:
[p eː n d͡ʒ]
* çînî:
[t͡ʃ iː n iː]
* karin:
[k aː ɾ ɪ n]
* nên:
[n eː n]
Occurrences:
184
Examples:
* bangê:
[b aː ŋ eː]
* renge:
[ɾ ɛ ŋ ɛ]
* lingê:
[l ɪ ŋ eː]
* mangê:
[m aː ŋ eː]

Stop
Occurrences:
846
Examples:
* pênc:
[p eː n d͡ʒ]
* pey.:
[p ɛ j]
* pêşîn:
[p eː ʃ iː n]
* pirsê:
[p ɪ ɾ s eː]
Occurrences:
2,249
Examples:
* sebrê:
[s ɛ b ɾ eː]
* babo:
[b aː b o]
* bûn?:
[b uː n]
* aborî:
[aː b o ɾ iː]
Occurrences:
2,361
Examples:
* qat:
[q aː t]
* tişt:
[t ɪ ʃ t]
* tirş:
[t ɪ ɾ ʃ]
* heftê:
[h ɛ f t eː]
Occurrences:
2,694
Examples:
* erd:
[ɛ ɾ d]
* dersa:
[d ɛ ɾ s aː]
* dor,:
[d o ɾ]
* doskî:
[d o s k iː]
Occurrences:
2,895
Examples:
* karin:
[k aː ɾ ɪ n]
* kalê:
[k aː l eː]
* çalak:
[t͡ʃ aː l aː k]
* lawik:
[l aː w ɪ k]
Occurrences:
831
Examples:
* goga:
[ɡ o ɡ aː]
* gîha,:
[ɡ iː h aː]
* begê:
[b ɛ ɡ eː]
* gori:
[ɡ o ɾ ɪ]
Occurrences:
496
Examples:
* qat:
[q aː t]
* eqilê:
[ɛ q ɪ l eː]
* ewqas:
[ɛ w q aː s]
* req:
[ɾ ɛ q]

Affricate
Occurrences:
538
Examples:
* çînî:
[t͡ʃ iː n iː]
* çalak:
[t͡ʃ aː l aː k]
* çeman:
[t͡ʃ ɛ m aː n]
* navçe:
[n aː v t͡ʃ ɛ]
Occurrences:
379
Examples:
* pênc:
[p eː n d͡ʒ]
* cem:
[d͡ʒ ɛ m]
* bicû.:
[b ɪ d͡ʒ uː]
* cemed:
[d͡ʒ ɛ m ɛ d]

Sibilant
Occurrences:
1,682
Examples:
* dersa:
[d ɛ ɾ s aː]
* doskî:
[d o s k iː]
* sebrê:
[s ɛ b ɾ eː]
* însan:
[iː n s aː n]
Occurrences:
915
Examples:
* fîrûz:
[f iː ɾ uː z]
* hozan:
[h o z aː n]
* zorî:
[z o ɾ iː]
* mêze:
[m eː z ɛ]
Occurrences:
1,120
Examples:
* êş,:
[eː ʃ]
* tişt:
[t ɪ ʃ t]
* tirş:
[t ɪ ɾ ʃ]
* pêşîn:
[p eː ʃ iː n]
Occurrences:
598
Examples:
* rojîn:
[ɾ o ʒ iː n]
* jin,:
[ʒ ɪ n]
* roja:
[ɾ o ʒ aː]
* hêja:
[h eː ʒ aː]

Fricative
Occurrences:
456
Examples:
* fîrûz:
[f iː ɾ uː z]
* heftê:
[h ɛ f t eː]
* fire:
[f ɪ ɾ ɛ]
* fanî:
[f aː n iː]
Occurrences:
1,115
Examples:
* ve,:
[v ɛ]
* navçe:
[n aː v t͡ʃ ɛ]
* viyan:
[v ɪ j aː n]
* çav,:
[t͡ʃ aː v]
Occurrences:
1,366
Examples:
* herî.:
[h ɛ ɾ iː]
* hozan:
[h o z aː n]
* heftê:
[h ɛ f t eː]
* rehme:
[ɾ ɛ h m ɛ]

Approximant
Occurrences:
1,061
Examples:
* xwe.:
[x w ɛ]
* lawik:
[l aː w ɪ k]
* wir?:
[w ɪ ɾ]
* dêw:
[d eː w]
Occurrences:
2,036
Examples:
* rûye:
[ɾ uː j ɛ]
* pey.:
[p ɛ j]
* yan:
[j aː n]
* dayik:
[d aː j ɪ k]

Tap
Occurrences:
4,718
Examples:
* karin:
[k aː ɾ ɪ n]
* rûye:
[ɾ uː j ɛ]
* rex:
[ɾ ɛ x]
* erd:
[ɛ ɾ d]

Lateral
Occurrences:
1,533
Examples:
* kalê:
[k aː l eː]
* çalak:
[t͡ʃ aː l aː k]
* lawik:
[l aː w ɪ k]
* mela:
[m ɛ l aː]
Occurrences:
13
Examples:
* millî:
[m ɪ ɫ iː]

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:
3,035
Examples:
* çînî:
[t͡ʃ iː n iː]
* doskî:
[d o s k iː]
* anîne:
[aː n iː n ɛ]
* fîrûz:
[f iː ɾ uː z]
Occurrences:
553
Examples:
* kurê:
[k u ɾ eː]
* kurên:
[k u ɾ eː n]
* kuçe:
[k u t͡ʃ ɛ]
* ruhê:
[ɾ u h eː]
Occurrences:
761
Examples:
* rûye:
[ɾ uː j ɛ]
* fîrûz:
[f iː ɾ uː z]
* rûnê:
[ɾ uː n eː]
* bûn?:
[b uː n]
Occurrences:
6,278
Examples:
* karin:
[k aː ɾ ɪ n]
* lawik:
[l aː w ɪ k]
* tişt:
[t ɪ ʃ t]
* dayik:
[d aː j ɪ k]
Occurrences:
1
Examples:
* özal:
[ʏ z aː l]

Close-Mid
Occurrences:
3,578
Examples:
* pênc:
[p eː n d͡ʒ]
* êş,:
[eː ʃ]
* kalê:
[k aː l eː]
* nên:
[n eː n]
Occurrences:
1,331
Examples:
* dor,:
[d o ɾ]
* doskî:
[d o s k iː]
* hozan:
[h o z aː n]
* babo:
[b aː b o]

Open-Mid
Occurrences:
7,836
Examples:
* rûye:
[ɾ uː j ɛ]
* pey.:
[p ɛ j]
* rex:
[ɾ ɛ x]
* xwe.:
[x w ɛ]

Open
Occurrences:
6,806
Examples:
* qat:
[q aː t]
* karin:
[k aː ɾ ɪ n]
* kalê:
[k aː l eː]
* çalak:
[t͡ʃ aː l aː k]
On this page
Show Source