New headphone system translates several speakers at once, clones voices in 3D sound

U.S. researchers have created a headphone system that translates multiple speakers simultaneously while preserving their voices' direction and qualities. The "Spatial Speech Translation" system uses noise-cancelling headphones with microphones to separate speakers, translate speech, and play it back with a short delay. It can detect multiple speakers, maintain voice characteristics, and track speakers as they move. The system currently works for common speech in Spanish, German, and French, with potential for expansion to 100 languages.