Americans and Australians are taking advantage of new features in Alexa with the ability to add playful or sad tones in the responses given by the voice assistant. Amazon also offers voice styles depending on the subject such as journalistic diction for reading information.
The Amazon developers have just released details of new features on their Alexa voice assistant on their blog . From now on, this artificial intelligence will be able to express different emotions, as well as adopt different styles for speaking.
To speak, Alexa uses technology speech neuronal Amazon, or Neural text-to-speech (NTTS). The first novelty, simply called Alexa Emotions , allows you to add an enthusiastic or sad tone, the two emotions having three degrees of intensity each. The personal assistant can thus be cheerful when the user correctly answers a question in a game, or sad when announcing that his favorite sports team has lost. The first tests showed that the satisfaction rate increased by 30% with the use of emotions.
New speech styles for news and music
Similarly, the developers have created new styles of voice. Users in the United States will be able to activate more natural styles for reading news like a journalist or for presenting music, while Australian users will be able to take advantage of journalistic style for now. local accent. To do this, they modify the intonation, the emphasis and the pauses in the sentence. The style for the news was judged more natural by 31% and the musical style by 84%.
Developers of Alexa voice applications will be able to integrate emotions and styles through the use of simple SSML tags around text, a markup language for speech synthesis based on XML. For now, these new products are only available in the United States and Australia. Amazon has not announced an availability date for other regions, but there is no doubt that this will happen soon with perhaps the diction of Laurent Delahousse or Claire Chazal for reading the news of the day.