Discover Adobe Firefly 2: Enhanced Vector Graphics, Advanced Control, and Stunning Photorealistic Renders

Adobe's Firefly generative AI, launched just seven months ago, is set to introduce three new models along with over 100 enhanced features, as announced at the Adobe Max 2023 event. The Firefly Image 2 model promises improved fidelity in generated images and offers users more precise control. Meanwhile, the new Vector model allows graphic designers to create vector images swiftly—making this a notable first in the industry. Additionally, the Design model introduces text-to-template generation for print and online advertising layouts, marking another pioneering advancement.

Adobe has a history of integrating machine learning into its products, with its first commercial AI, Sensei, being released in 2016. Built upon the Sensei framework, Firefly provides a comprehensive suite of AI tools designed for image and video editing. This includes features like “text to color enhancement,” tools for adjusting saturation and hue, font and design element generation, and even the ability to create and add background music to videos seamlessly. The Firefly suite is accessible across Adobe’s range of products, including Premiere Pro, After Effects, Illustrator, Photoshop, and Express, available on all subscription tiers of the Creative Cloud platform, including the free plan.

The updated Firefly Image 2 model retains its predecessor's focus on licensed and public domain content, ensuring that its generated images are safe for commercial use. It accommodates text prompts in 100 languages. The Firefly AI operates across various mediums—still images, videos, audio, design elements, and font effects. The new Firefly Vector model, currently in beta, adds the capability to generate vector art. This model features Generative Match, which enables users to replicate specific artistic styles in their outputs, facilitating brand consistency and rapid design creation utilizing existing aesthetics, as well as seamless, tileable fill patterns and vector gradients.

The Design model is tailored for advertising and marketing professionals, enabling them to generate print and online copy templates using Adobe Express. Users can create images in Firefly and then easily transfer them to Express for layout generation based on their natural language prompts. These templates are adaptable and can be created in popular aspect ratios.

The Firefly web application will also introduce three new features: Generative Match, which ensures consistent design aesthetics; Photo Settings, which aims for more photorealistic images and allows users to adjust photography metrics like depth of field and blur; and Prompt Guidance, which refines user input to enhance the effectiveness of image generation. With these advancements, Firefly continues to enhance user experience while streamlining the creative process.