A real-time voice translator chains three steps: speech recognition turns spoken words into text, machine translation converts that text into another language, and speech synthesis reads it aloud. One maker built such a setup to hold a conversation with a French speaker. The original video is no longer available.