top of page
  • Writer's pictureJerry Garcia

Adobe Unveils Groundbreaking Text-to-Video Generative AI Tools

Adobe has unveiled a preview of its upcoming generative AI video tools, showcasing features that can transform text descriptions and still images into video clips. This innovation is part of Adobe's Firefly video model, which aims to revolutionize video and audio editing within Adobe’s Creative Cloud applications.

Key Takeaways

  • Adobe's new AI tools can generate video clips from text descriptions and still images.

  • The Firefly video model promises high-quality, commercially safe content.

  • Initial beta release expected later this year.

Firefly's Text-to-Video Capabilities

Adobe's Firefly video model introduces groundbreaking text-to-video capabilities. Users can generate video clips by simply inputting text descriptions. The tool also offers various “camera controls” to simulate different angles, motion, and shooting distances. This feature is expected to be particularly useful for creating additional B-roll footage or filling gaps in production timelines.

Image-to-Video Feature

In addition to text-to-video, Adobe demonstrated an image-to-video feature. This allows users to generate video clips using specific reference images. The generated video quality is comparable to OpenAI’s Sora model, although Adobe's VP of generative AI, Alexandru Costin, noted that the maximum length for these videos is currently limited to five seconds.

Commercial Safety and Integration

One of the standout features of Adobe's Firefly model is its promise of being “commercially safe.” The model is trained on openly licensed, public domain, and Adobe Stock content, which helps mitigate concerns about copyright infringement. Adobe plans to integrate these new tools into its Creative Cloud, Experience Cloud, and Adobe Express applications.

Generative Extend Feature

Adobe also previewed the “Generative Extend” feature for Premiere Pro. Similar to Photoshop’s Generative Expand tool for image backgrounds, this feature can extend the length of existing video footage. Adobe has not yet specified a release date for this feature, but it is expected to arrive later this year.

Future Prospects

The text-to-video and image-to-video features will initially be available in beta as a standalone Firefly application. Adobe's advancements in generative AI video tools are set to offer new creative possibilities for content creators, making video production more efficient and versatile.

Stay tuned for more updates as Adobe continues to innovate in the realm of generative AI.

Sources

  • Adobe previews its upcoming text-to-video generative AI tools - The Verge, The Verge.

  • Adobe - Adobe Previews Its Upcoming Firefly Video Model, Adobe.

0 views

Comments


bottom of page