Generate Videos from Text Using Google Veo 2
Since the introduction of Gemini, Google has mainly concentrated on adding features like image generation and various model integrations. Now, the company is enhancing its capabilities in video generation by expanding access to the Veo 2 video generator through its AI chatbot.
The Veo 2 was first announced in December of the previous year, which promised a higher level of realism in video production. It features improved rendering techniques that better mimic real-life physics, human movements, and intricate details compared to earlier models. Google has now unveiled the integration of the Veo 2 generator into the Gemini platform, making it available to Gemini Advanced users who subscribe to the Google One AI Premium service on both mobile and web platforms.
Generating Videos with Gemini
Users can create videos by selecting the Veo 2 model from the AI model selection menu and then inputting their text prompts. Google suggests that detailed prompts lead to more accurate video outputs. This new capability supports a range of styles, including cinematic effects and various film genres.
Step-by-Step Guide to Video Generation
- Select Veo 2 from the model selection menu.
- Enter a detailed text prompt.
- Note that video generation is capped per month (exact limits are yet to be defined).
Each video generated will last for 8 seconds and will be in 720p resolution with a 16:9 aspect ratio. This is notably shorter and at a lower resolution than similar offerings from competitors, like OpenAI’s Sora, which can create videos up to 20 seconds long in 1080p quality. The videos produced will be in MP4 format, allowing users to save them easily. Additionally, mobile users will find features to share their generated videos directly on popular social media platforms like YouTube and TikTok.
Digital Watermarking
Google also incorporates digital watermarks into the videos created with Veo, known as SynthID watermarks. This is similar to the watermarking in AI-generated images from Gemini or the Imagen model, ensuring authenticity and credit for the generated content.
Extending Functionality with Whisk
In addition, Google is expanding the functionality of its experimental feature, Whisk. This application allows users to create images using both text and images. Now, with the integration of Veo 2, users can bring their static images to life, generating animated versions of the images they create in Whisk. These animations will also last for 8 seconds and remain in the MP4 format, matching the output from Gemini.
Availability of Veo 2
The rollout of the Veo 2 model is currently underway for Gemini Advanced users and is being offered in English. Users may experience a delay of up to a day before the model becomes available to all subscribers.
Have you tried any AI video generators? Which one do you prefer? Feel free to share your experiences in the comments below.
