New Google tool creates video with text

Imagen Video uses AI to generate creative videos from a written prompt using extensive video and picture collection

Google launched a text-to-video artificial intelligence (AI) tool, called Imagen Video, right after Meta released its own tool "Make a Video".

The tool will be able to produce videos at 24 frames per second in 1280x768, entirely from a written prompt. With this tool in beta mode, users can create fantasy buildings and Pixar-style monsters. A demo of Google's tool had included coffee being poured into a cup, reported Business Today.

Read: Elon Musk's SpaceX buys ad campaign on Twitter for Starlink

The features of Imagen Video include generated videos of artwork by famous artists, 3D rotating objects, and rendered texts through different animation styles. Moreover, the tool would include noise cancellation augmentation and an internal data set of 14 million videos and 60 million images.

How does it work?

Users will have to input text prompts, which will be encoded into textual embedding, and will then generate a 16-frame video. The video would then be used to create a 128-frame video with a high definition of 5.3s.

While other companies like Meta, Microsoft, and DALL-E2 have worked on similar video creators, Google's Imagen Video has been criticised for copyright issues, abusing artwork and posing a threat to future artists.