Amazon says that its AI video model can now generate several minutes clips
Amazon has improved his AI video model, Nova ReelWith the ability to generate videos for up to two minutes.
Nova Reel, announced in December 2024, was Amazon’s first foray into the generative video space. He competes with the models of OPENAI,, GoogleAnd others in what quickly becomes a crowded market.
The latest NOVA coil, Nova Reel 1.1, can generate “multi-shot” videos with “coherent style” through shots, explained that the developer AWS defended Elizabeth Fuentes in a blog article. Users can provide a prompt up to 4,000 characters to generate up to a two -minute video made up of six seconds.
Nova Reel 1.1 also presents a new mode called “Multinist Manual”. In this mode, the model can refer to an image with an prompt to offer more control over the composition of a video photo. According to Fuentes, given an image 1280 × 720-resolution and invites it 512-maximum-caractre, the multi-stroke manual can generate videos containing up to 20 shots.
Nova Reel is only available via AWS platforms and services, including the foundation, AMAzon AI development suite and customers must request special access. As for most generative AI systems, there are questions as to Reel has been developed using ethical methods.
Video generating models are formed on a large number of examples of videos to “learn” the models of these videos to generate new clips. Some companies form models on videos protected by copyright without obtaining permission from owners or creators, and, when these models “regurgitateSabile protected by copyright, he exhibits users of models to IP prosecution.
Amazon has not revealed the source of Reel training data, and it did not provide an explicit means for creators to supply the data sets of the models with their videos to withdraw. However, the company said that it will protect all AWS customers accused of having raped copyright with the media generated by its models, in accordance with its compensation policy.
