This is a dataset created for academic research in voice activation. The format of dataset is the same as in Google Speech Commands dataset.
In raw directory you can see:
- original wav files (1 channel, 16 bit, 16 khz)
- text files with segmented words
- cut.py - python script for preparing dataset from wav files
- words - list of words in the same order as in recordings
In dataset directory you can see the dataset itself.