Extended Data
- Audio 1, WAV file -
Reconstructed audio files of a selection of spoken words and environmental sounds, based on the convolution model and spectrogram. The original stimulus is presented first, followed by the reconstructed sound; this order of presentation is used to ease the listener’s perception of relevant speech features, but please note that it produces a priming effect.
- Audio 2, WAV file -
Reconstructed audio files of a selection of environmental sounds, based on the convolution model and spectrogram. The original stimulus is presented first, followed by the reconstructed sound.