Amazon has launched two new additions to its generative AI portfolio—Amazon Nova Sonic, a basis mannequin for voice-based purposes, and Amazon Nova Reel 1.1, an up to date mannequin for text-to-video technology.
Nova Sonic integrates speech recognition, understanding and technology into one mannequin, eradicating the necessity for separate elements. Conventional voice techniques contain complicated pipelines—changing speech to textual content, processing by way of a big language mannequin, and changing the response again to speech. In line with Amazon, this strategy “fails to protect essential acoustic context and nuances.”
“Nova Sonic takes a brand new strategy,” the corporate stated. “It unifies the understanding and technology capabilities right into a single mannequin.” The result’s a voice agent that not solely understands person enter but in addition responds with an applicable tone, tempo, and elegance.
The mannequin is out there by way of Amazon Bedrock. It helps purposes in customer support, journey, schooling, healthcare and leisure. In a single instance shared by Amazon, a digital journey assistant shifts its tone in response to a buyer’s change in emotion—transferring from enthusiastic to reassuring when issues about value are raised. One other use case contains an enterprise dashboard assistant that grounds solutions in firm knowledge and maintains multi-turn dialogue with out requiring customers to reset context.
Nova Sonic additionally generates transcripts of person speech. This function permits builders to combine exterior APIs and instruments, enabling AI brokers to carry out duties equivalent to retrieving flight choices or accessing inner dashboards.
Then again, Nova Reel 1.1 allows multi-shot movies as much as two minutes in size, with constant visible model throughout 6-second segments. It improves on the earlier model when it comes to technology pace and coherence. Customers can select to offer a single immediate for your complete video or set particular person prompts per shot for extra management.
The mannequin helps use circumstances like advertising campaigns, product design showcases, and social media content material creation. “Nova Reel enhances inventive productiveness,” Amazon stated, “whereas serving to to scale back the time and value of video manufacturing utilizing generative AI.”
To get began with Amazon Nova Reel 1.1, customers want to go to the Amazon Bedrock console and request entry to the mannequin. Within the left-hand navigation panel, they need to choose “Mannequin entry” after which find Amazon Nova Reel within the listing of accessible fashions. Requesting entry right here gives permission to make use of each model 1.0 and 1.1 of the mannequin. As soon as entry is granted, customers can start utilizing Amazon Nova Reel 1.1 by way of the Amazon Bedrock console, the AWS SDK, or the AWS Command Line Interface (CLI).
The releases are a part of Amazon’s broader Nova mannequin household, launched at re:Invent 2024, which additionally contains Nova Micro, Lite, and Professional that generate textual content from totally different modalities
The publish Amazon Rolls Out Nova Sonic and Nova Reel 1.1 for Generative Voice and Video AI appeared first on Analytics India Journal.
