Google's Veo 3.1 video model upgrade

Veo 3.1, a new video generation model just released by Google, promises enhanced realism, quality enhancements, improved image-to-video capabilities, and a number of new editing tools aimed at filmmakers and creative control. The specifics: In order to preserve character coherence throughout scenes, Veo 3.1 now supports up to three reference photos. Additionally, users can supply start and end frames, and 3.1 will produce seamless transitions between them along with corresponding sounds. By repeatedly adding parts that match the preceding clip, users may now build videos up to one minute long thanks to new scene extension capabilities. Both the normal and fast versions of 3.1 are being released throughout Google's ecosystem, which includes Gemini, Vertex AI, and its Flow filmmaking tool. Despite what the benchmarks may indicate, Veo 3.1 doesn't have the same buzz as Sora 2, which just weeks earlier set the standard for AI videos in a way that went viral. The greater improvement might be in the editing area, where features like start/end frames and scene stretching provide the additional control required to elevate outputs.