Example

Audio Samples

Below are paired audio samples from the TAPS dataset. Each pair contains recordings from both acoustic and throat microphones for the same speech segment. The acoustic microphone captures normal speech, while the throat microphone captures vibrations directly from the vocal cords.

Accessing the Full Dataset

The complete TAPS dataset is available on Hugging Face and can be loaded easily with a few lines of code:

from datasets import load_dataset
dataset = load_dataset("yskim3271/Throat_and_Acoustic_Pairing_Speech_Dataset")

# Access a sample
sample = dataset["train"][0]
throat_audio = sample["audio.throat_microphone"]
acoustic_audio = sample["audio.acoustic_microphone"]