Download Files and Resources¶
Perceptimatic gathers small datasets built in various projects. All of them were made from ABX tests on humans. Each stimulus is a triphone and usually the second sound is tested on the subject. A short reminder of their contents is available in the following table. For more information on the subset sources and contents, please refer to section Dataset description):
Note
When using the Perceptimatic dataset for publications, please cite our paper on arxive PAPER SUMMARY OF THE ALL DATASET ?
Perceptimatic Dataset Files¶
Sound files¶
Subset Name |
Description |
Source Wav Files |
Extracted wav files |
Triplets Wav Files |
|---|---|---|---|---|
CogSci 2019 |
French and English cross-language ABX stimuli triplets tested on french and english subjects |
|||
Zero Resource Speech Challenge 2017 |
English and French ABX stimuli triplets tested on french and english subjects |
|||
Pilot Dataset July 2018 |
ABX stimuli triplets from various English dialects, extracted from TIMIT dataset (English dialects) |
|||
Pilot Dataset August 2018 |
English ABX stimuli triplets |
|||
WorldVowels dataset |
Multilingual ABX stimuli triplets |
Annotation files¶
Subset Name |
Description |
Human experiments |
Stimuli meta information |
Triplets Meta information |
|---|---|---|---|---|
CogSci 2019 |
French and English cross-language ABX stimuli triplets tested on french and english subjects |
|||
Zero Resource Speech Challenge 2017 |
English and French ABX stimuli triplets tested on french and english subjects |
|||
Pilot Dataset July 2018 |
ABX stimuli triplets from various English dialects, extracted from TIMIT dataset (English dialects) |
|||
Pilot Dataset August 2018 |
English ABX stimuli triplets |
|||
WorldVowels Dataset |
Multilingual ABX stimuli triplets |