Adobe Firefly Image 5: Layers & Custom Models Arrive

Adobe Launches Firefly Image 5 with Enhanced Capabilities

On Tuesday, Adobe announced the release of Firefly Image 5, the newest version of its powerful image generation model.

Alongside this update, Adobe is expanding the functionality of the Firefly website, increasing support for models from other AI developers, and introducing features for generating both speech and soundtracks.

Key Improvements in Firefly Image 5

A significant advancement is the ability for artists to develop their own personalized image models utilizing their pre-existing artwork.

Image 5 now operates with native resolutions reaching up to 4 megapixels. This represents a substantial leap forward compared to the previous generation, which generated images at 1 megapixel and then upscaled them.

The company also reports improved rendering of human subjects within the generated images.

Layered Editing and Precise Control

Image 5 introduces layered and prompt-based editing capabilities.

The model now recognizes distinct objects as individual layers, allowing for targeted modifications through prompts. Tools for resizing and rotating elements are also included.

Adobe emphasizes that these editing processes are designed to preserve the image’s overall detail and structural integrity.

Expanding the Ecosystem with Custom Models

Adobe’s Firefly platform already supports third-party models from leading AI labs, including OpenAI, Google, Runway, Topaz, and Flux.

Now, users can create custom models reflecting their unique artistic styles. This feature, currently in closed beta, allows users to upload assets like images, illustrations, and sketches.

Redesigned Firefly Website for Streamlined Workflow

The Firefly website, recently redesigned, offers new features for a more efficient user experience.

Users can now seamlessly switch between image and video generation using the prompt box, select preferred AI models, and adjust aspect ratios.

The homepage now displays recent files and generation history, with quick access to other Adobe applications.

Enhanced Video Tools with Layer Support

Adobe has revamped the video generation and editing tool to incorporate layers and timeline-based editing.

This update is currently available in a private beta and will be released to all users in the future.

New Audio Features: Soundtracks and Speech Generation

Firefly is gaining two new audio capabilities.

Users can now generate complete soundtracks and speech for videos using AI prompts and models from ElevenLabs.

A new prompt creation method is also available, utilizing keywords and sections selected from a dynamic word cloud.

Adobe's Strategy for the Future of Creativity

As competitors like Canva integrate AI into their platforms, Adobe is focusing on attracting a new generation of creators.

According to Alexandru Costin, VP of generative AI at Adobe, the target audience is “creators or next-generation creative professionals” who readily embrace GenAI in their workflows.

Costin notes that Firefly’s development allows for greater flexibility in adding new features and experimenting with the interface, free from the constraints of established workflows in Adobe’s Creative Cloud suite.

Firefly represents Adobe’s commitment to innovation in the evolving landscape of AI-powered creative tools.