Google has announced enhancements to its AI filmmaking tool, Flow, aimed at increasing the realism of videos generated through the platform. The updated features allow users to adjust shadows and lighting within their AI videos, contributing to a more lifelike appearance that may be difficult to identify as AI-generated. These improvements are part of the Veo 3.1 update, which Google also introduced concurrently.
The Flow tool now supports audio integration, allowing users to create videos with sound. This functionality can be achieved through new features such as “Ingredients to Video,” where users can generate audio-visual content based on three reference images. Another capability, known as “Frames to Video,” connects a starting image to an ending image, accompanied by audio. Additionally, users can utilize the “Scene Extension” feature to append up to one minute of generated video to the final second of an existing clip, complete with generated audio.
The Veo 3.1 update retains the pricing structure of its predecessor, Veo 3, and is currently available as part of a paid preview through the Gemini API for developers. It is also enabled in the Gemini app.
Furthermore, Google stated that Flow users will soon have the ability to remove any object from a video. This feature will restructure the background and scene to create the illusion that the removed object was never part of the video. The ongoing developments raise questions about the implications of such advancements in AI-generated content and their potential impact on media authenticity.
Source: https://www.theverge.com/news/800371/google-veo-3-1-flow-audio