Speaking to a customer service rep with a thick accent is difficult enough; when the connection is bad, it's even worse. Similarly, even for those who speak multiple languages, understanding someone speaking in a different language in a noisy environment, like a transportation hub or a convention center, can be challenging.

Chinese electronics manufacturer Timekettle has designed their W4 AI Interpreter Earbuds for the latter application. They use bone conduction technology (in addition to a microphone) to accurately capture speech even in noisy environments. They need to be paired with your phone to connect to the company's cloud translation service, and as you speak, your translated communications are played through your phone's speaker—in your voice, cloned by AI.

That said, listening to a phone speaker in a noisy environment isn't ideal. To get around this, the company has developed a One-on-One mode, which comes with a slight ick factor: You give one earbud to the person you're speaking with, so you've each got one. When you speak, the other party hears the translation directly through the earbud, and vice versa.

The company says that the translation time is just 0.2 seconds (I imagine it actually depends on your connection speeds), and that they can handle 42 languages and 95 accents with "up to 98%" accuracy.

The buds offer up to four hours of translation time on a single charge. They run $350.