|
|
--- |
|
|
language: |
|
|
- ab |
|
|
- am |
|
|
- ba |
|
|
- be |
|
|
- bg |
|
|
- bn |
|
|
- ca |
|
|
- cs |
|
|
- cv |
|
|
- ckb |
|
|
- dv |
|
|
- el |
|
|
- eu |
|
|
- gn |
|
|
- ha |
|
|
- hi |
|
|
- hsb |
|
|
- hu |
|
|
- hy |
|
|
- id |
|
|
- it |
|
|
- ja |
|
|
- ka |
|
|
- kk |
|
|
- ko |
|
|
- ky |
|
|
- lt |
|
|
- mk |
|
|
- mn |
|
|
- mr |
|
|
- mt |
|
|
- nl |
|
|
- or |
|
|
- pa |
|
|
- pl |
|
|
- pt |
|
|
- ro |
|
|
- ru |
|
|
- rw |
|
|
- sk |
|
|
- sl |
|
|
- sq |
|
|
- sr |
|
|
- sv |
|
|
- sw |
|
|
- ta |
|
|
- th |
|
|
- tk |
|
|
- tr |
|
|
- ug |
|
|
- uk |
|
|
- uz |
|
|
- vi |
|
|
- yo |
|
|
- yue |
|
|
- zh |
|
|
tags: |
|
|
- Phonetics |
|
|
- Linguistics |
|
|
- Corpus |
|
|
size_categories: |
|
|
- n<1K |
|
|
pretty_name: VXC |
|
|
license: cc0-1.0 |
|
|
--- |
|
|
## VoxCommunis Corpus |
|
|
The VoxCommunis Corpus is a phonetic corpus derived from the Mozilla Common Voice Corpus. Corresponding audio files and corpus metadata can be downloaded from [Mozilla Common Voice](https://commonvoice.mozilla.org/en), or from one of several Hugging Face repositories for the differing versions. |
|
|
|
|
|
Within each folder, the filenames share similar structure and contain critical information for effectively using the file. More detail regarding the specifics of the filename for each file type is provided below. In general, a filename such as ``mk_xpf_lexicon19`` corresponds to: |
|
|
|
|
|
+ ``mk``: Common Voice language ID code (Macedonian) |
|
|
+ ``xpf``: G2P system for the lexicon (XPF Corpus) |
|
|
+ ``19``: Common Voice version (Macedonian Version 19) |
|
|
|
|
|
``acoustic_models/``: The acoustic models have been trained using the Montreal Forced Aligner, and the force-aligned TextGrids are obtained directly from those alignments. These acoustic models can be downloaded and re-used with the Montreal Forced Aligner for new data. |
|
|
|
|
|
``lexicons/``: The lexicons are developed using various toolkits. Some manual correction has been applied, and we hope to continue improving these. Any updates from the community are welcome. |
|
|
|
|
|
+ ``epi``: Epitran |
|
|
+ ``xpf``: XPF Corpus |
|
|
+ ``chr``: Charsiu |
|
|
+ ``cvu``: Common Voice Utils |
|
|
+ ``mfa``: Montreal Forced Aligner G2P |
|
|
+ ``vxc``: Custom dictionaries |
|
|
|
|
|
``textgrids/``: The TextGrids contain phone- and word-level alignments of the validated set of the Common Voice data. A filename such as ``mk_xpf_textgrids19_acoustic19`` corresponds to: |
|
|
|
|
|
+ ``mk``: Common Voice language ID code (Macedonian) |
|
|
+ ``xpf``: G2P system for the lexicon (XPF Corpus) |
|
|
+ ``textgrids19``: Common Voice version for the TextGrids (Macedonian Version 19) |
|
|
+ ``acoustic19``: Common Voice version for acoustic model (Macedonian Version 19) |
|
|
|
|
|
``similarity_scores/``: The client ID similarity scores generated using the automatic speaker verification procedure described in Zhang, M., Farhadipour, A., Baker, A., Ma, J., Pricop, B., Chodroff, E. (2025) Quantifying and Reducing Speaker Heterogeneity within the Common Voice Corpus for Phonetic Analysis. Proc. Interspeech 2025, 3933-3937, doi: 10.21437/Interspeech.2025-2027. More details can be found at [pacscilab/CV_clientID_cleaning](https://github.com/pacscilab/CV_clientID_cleaning). |
|
|
The filenames have the structure of: {language ID}_{version number). The files have the column structure: |
|
|
|
|
|
+ enroll: The enrollment filename |
|
|
+ test: The test filename |
|
|
+ score: The cosine similarity score generated by the automatic speaker verification system |
|
|
|
|
|
``speaker_files/``: The spkr_files are based on the data contained in ``validated.tsv`` included in the Common Voice download. They contain one additional mapping from the original client_id to a simplified spkr_id. |
|
|
|
|
|
``similarity_scores_interspeech25.txt``: The similarity scores from 76 languages analyzed and reported in Zhang, M., Farhadipour, A., Baker, A., Ma, J., Pricop, B., Chodroff, E. (2025) Quantifying and Reducing Speaker Heterogeneity within the Common Voice Corpus for Phonetic Analysis. Proc. Interspeech 2025, 3933-3937, doi: 10.21437/Interspeech.2025-2027 |
|
|
|
|
|
### Additional code and resources |
|
|
The primary Github repository for the VoxCommunis Corpus can be found here: [https://github.com/pacscilab/voxcommunis](https://github.com/pacscilab/voxcommunis) |
|
|
|
|
|
The Github repository for the Common Voice client ID auditing can be found here (with links for additional code contained therein): [https://github.com/pacscilab/CV_clientID_cleaning](https://github.com/pacscilab/CV_clientID_cleaning) |
|
|
|
