Sculpted Speech

This dataset contains the stimuli heard by participants in a study of sculpted speech. The zip file contains 240 Spanish sentences from the Sharvard corpus in each of eight conditions that vary in the amount of target speech information they contain.  

For illustration purposes (for those not wanting to download the entire zip), the dataset also contains as separate wav files for one example sentence in each of the eight conditions. The names for these wav files map to the subdirectories as follows:

baseline -> 'mix'
speech -> 'mask_speech'
mix -> 'mask_mix'
fine -> 'fine_mix'
envelope -> 'fine_masker'
ssn -> 'mask_ssn'
music -> 'mask_music'
compspeech -> 'mask_babble'