Google's new VideoPoet AI tool can generate videos based on text input

About a year ago, we were astonished by websites that could generate original images based on the user’s text input. Today, Google has introduced something similar yet surprisingly novel.

Google has announced a new AI tool called VideoPoet that can generate videos based on text input. The tool can create various types of videos, including short films, music videos, and even explainer videos.

VideoPoet is a large language model trained on a massive text and video data dataset. As a result, the model is capable of comprehending the relationship between text and video and can produce videos that are both coherent and visually appealing.

One of the most impressive things about VideoPoet is its ability to generate long videos. The model can create videos that are several minutes long by chaining together short clips. This makes it possible to use VideoPoet to create more complex and nuanced videos.

VideoPoet can also be used to edit existing videos. For example, you can use the tool to add animation to a static image or change a video’s style. This makes VideoPoet a powerful tool for video editing, even for people who don’t have a lot of experience.

Here are some of the specific tasks that VideoPoet can be used for:

• Text-to-video: Generate a video from a textual description.
• Image-to-video: Animate a static image.
• Video stylization: Apply a stylistic effect to a video.
• Video inpainting and outpainting: Fill in missing parts of a video or extend a video beyond its original boundaries.
• Video-to-audio: Generate audio from a video clip.

Google is still working on VideoPoet, but the tool can already produce impressive results. You can try it here.