OpenAI’s Sora has opened the way to AI-powered video-generating apps that create short clips from text prompts. This week, Google announced its own video-generating AI app, Google Veo, at its annual I/O Developer Conference.
As with many other recently announced AI tools, the new video generator is not yet available to the public. Artists and creators will be able to try out the tool through crowdsourced exercises, which can help it improve. Here are some other things we learned about Google Veo.
What Can Google Veo Do?
Google Veo can produce 1080p videos over a minute long and can also be edited through written instructions. Users can use a single prompt or a series of prompts to create a narrative, and frames can be visually consistent. Aerial shots and highly detailed scenes can be created with the tool.
When creating videos with Veo, you can choose between realism, surrealism, and even animation. It also understands terms such as ‘timelapse’ and ‘aerial shot’ and can use filmmaking commands. Veo offers greater control and superior video quality than previous AI video generators announced by Google.
Masked Editing allows you to adjust specific sections of the video with Veo. The prompt can include a mask area, and images can also be used with text prompts to generate videos.
This video was generated using an image of alpacas wearing knitted sweaters and the prompt Alpacas dancing to the beat as a reference.
How Does Google Veo Work?
Video AI models must be able to comprehend the prompts correctly, and combine that information with references available. Due to the way Veo understands natural language and visuals, it produces accurate videos, including those with several details.
This is why you can produce a high-quality video using a prompt like Timelapse of a common sunflower opening, dark background.
According to Google, Veo is a video-generating tool that improves upon Lumiere, VideoPoet, WALT, etc., by utilizing compressed latent video utilizations. With Veo, the AI understands prompts better since it is trained on detailed video captions.
According to Google, the video generator uses latent diffusion transformers, which ensure consistency and prevent characters and objects from flickering. The video is also compressed using latents, which help improve performance and efficiency. As a result, Veo can produce high-quality videos more quickly.
The videos you create using Veo will be watermarked using Google’s SynthID, which uses the Coalition for Content Provenance and Authenticity (C2PA) standard to identify AI-generated content.
How Can You Use Google Veo?
Currently,Google Veo is not available for public use, and Google is only letting a few select creators use it. VideoFX is another experimental tool available in Google Labs, where all of Google’s in-development products can be tested.
You can try Veo if you are a creator
- Visit the Google Labs website to join the waitlist.
- Enter your details and submit them.
- Your submission will be checked by Google. A Veo trial will be allowed if it is found suitable.
Google Veo is the latest in a long line of AI-powered video-generators announced by Google, and it seems to be the most powerful one so far. You can create videos using different styles and effects, as well as edit them.
We’ve explained how creators can join the waitlist and try out Veo for themselves, but we don’t know when it will roll out publicly. Some of the tool’s capabilities may make their way into YouTube shorts, so there’s that to look forward to.
