🗣️ DeepL's new Aure model translates your voice live, sarcasm and mood swings included
DeepL has unveiled Aure, a new model powering DeepL Voice that carries a speaker's changing tone, emotion, emphasis and pace into the translated speech in real time. The old system set the translated voice from a snapshot at the start of a call, so it could not follow a mood shift or sarcasm; Aure keeps analyzing the source audio and the translated words as the conversation goes. It runs in DeepL Voice for meetings, in-person conversations and the API, starting with eight languages including English, Japanese and Korean.
⭐ Live translation is moving from getting the words right to sounding like the actual person, a key step for business calls across languages.
#AI
DeepL has unveiled Aure, a new model powering DeepL Voice that carries a speaker's changing tone, emotion, emphasis and pace into the translated speech in real time. The old system set the translated voice from a snapshot at the start of a call, so it could not follow a mood shift or sarcasm; Aure keeps analyzing the source audio and the translated words as the conversation goes. It runs in DeepL Voice for meetings, in-person conversations and the API, starting with eight languages including English, Japanese and Korean.
⭐ Live translation is moving from getting the words right to sounding like the actual person, a key step for business calls across languages.
#AI
