In our Veo 3.1 AI video model, Google pays close attention to the reference images that the generated clips are based on. In addition to native vertical video support and expanded resolution upscaling capabilities, the company is releasing new visual improvements to its “Ingredients to Video” tool introduced last year.
The Elements to Video tool allows Veo users to generate videos based on up to three reference images and bring in materials such as character subjects, backgrounds, and textures for more control over how the results look. Google says the update will make videos “more expressive and creative” and offer “richer interactions and storytelling.” There are also more noticeable consistency improvements. With Veo 3.1, characters will now look the same across different clips and environments, allowing users to reuse objects, backgrounds, and textures between scenes.
Clips generated using Elements to Video now also support vertical output. This follows last year when Google gave developers the ability to generate vertical videos in Veo for unreferenced text-based prompts. Instead of manually cropping the results in video editing software, users can choose to output videos in a native 9:16 aspect ratio that can be uploaded to platforms like TikTok and YouTube Shorts.
Starting today, Google is adding Veo’s improved Individual to Video and Portrait Mode features to its Gemini app, and integrating these tools into its YouTube Shorts and YouTube Create apps “for the first time.”
Finally, this update allows Veo 3.1 users to upscale their generated videos from the previous 1080p limitation to 4K resolution. 1080p video generation has also been improved to provide “sharper, cleaner video,” according to Google. This isn’t the native 4K resolution that Google claimed Veo would be able to produce in 2024 (it hasn’t been confirmed in any publicly available versions of Veo yet), but upscaling on the platform is better than nothing.
