Dataset · Cnam-LMSSC
French speech recorded with bone conduction sensors
Cnam-LMSSC/vibravox
Audio dataset of French speech captured by body conduction sensors for recognition, synthesis, and audio enhancement tasks.
What's inside
The dataset contains between 10K and 100K French speech recordings and corresponding texts obtained using specialized body (bone) conduction sensors. This unusual recording method makes the data useful for audio processing in complex acoustic environments.
What you can do with it
- Training French automatic speech recognition (ASR) and text-to-speech (TTS) systems
- Training audio-to-audio conversion and enhancement models using bone conduction sensors
- Tasks
- Audio → audio, Automatic speech recognition, Audio classification, Text → speech
- Languages
- French
- Size
- 10K–100K rows
- Downloads this month
- 12.4K
- Likes
- 31
Models trained on it
- Cnam-LMSSC/phonemizer_headset_microphoneAutomatic speech recognition 139
- Cnam-LMSSC/EBEN_throat_microphoneAudio → audio 80
- Cnam-LMSSC/EBEN_temple_vibration_pickupAudio → audio 59
- Cnam-LMSSC/EBEN_forehead_accelerometerAudio → audio 40
- Cnam-LMSSC/EBEN_soft_in_ear_microphoneAudio → audio 34
- Cnam-LMSSC/EBEN_reverse_temple_vibration_pickupAudio → audio 34
- Cnam-LMSSC/EBEN_rigid_in_ear_microphoneAudio → audio 33
- Cnam-LMSSC/EBEN_noisy_soft_in_ear_microphoneAudio → audio 31
How to load it
from datasets import load_dataset
ds = load_dataset("Cnam-LMSSC/vibravox")