The University of Washington has presented innovative AI headphones that can simultaneously translate the speech of several speakers. At the same time, they preserve voice features and the direction of sound in space.

The device is powered by the Apple M2 chip. It uses a Sonic Presence SP15C binaural microphone and the housing from Sony WH-1000XM4 headphones to process audio in 2-4 seconds.

The headphones detect the number of speakers and their location, and translate from Spanish, German or French with high accuracy. All processes take place locally, which ensures the privacy of user data. The technology supports spatial sound, creating an effect of presence and adapting the volume depending on the distance to the speaker.

It is still unclear whether the headphones will become a commercial product. However, the project code is open to developers, which may contribute to further development of the technology