At the MAX conference in London, Adobe unveiled the newest version of their image generation model Firefly, just a week after we got new Generative AI features in Premiere Pro.
Highlights:
- Adobe released the upgraded Firefly Image 3 model with better image generation capabilities.
- The new model supports reference image generation and a better generative Expand feature in the Generative Fill module.
- The overall user feedback on this model has been moderate, with many stating that Midjourney v6 is still better.
Firefly Image 3 is Here
Adobe Firefly Image 3 is the latest version of the company’s image generation model with better photorealistic quality, the ability to understand longer prompts and text generation.
It is available in the Firefly web app, Adobe Photoshop, and Adobe InDesign in beta.
Here’s what they told us in the official announcement for the new launches at MAX Conference:
“Firefly Image 3 Foundation Model will bring massive advancements to quality and control to Firefly’s foundation image model. This includes higher-quality image generations, a better understanding of prompts, new levels of detail and variety, and significant improvements that enable fast creative expression and ideation. Adobe’s newest Firefly model delivers photorealistic quality like never before with better lighting, positioning, attention to detail, as well as advancements in text display.”
This new model makes it easier to use Photoshop’s AI tools plus an upgraded image generation model that offers higher quality outputs, better prompt understanding, and more detailed images.
Adobe’s previous attempts at developing an image generation model have rather fallen flat. In October 2023 when they released image 2, many users noted that compared to advanced frameworks like Midjourney, the images generated by Firefly seemed “stock-image” quality. The third iteration of the model is predicted to perform much better than its predecessors.
Like its previous versions, Image 3 is trained on Adobe Stock, Adobe’s royalty-free media library, along with licensed and public domain content for which the copyright has expired.
Adobe presents itself as the more ethical option over other image generators since Adobe pays the contributors to its training data.
New Features of Firefly Image 3
The standout feature of Firefly Image 3 is the Reference Image tool, which allows users to influence the AI-generated output by providing a reference image.
Adobe’s AI analyzes the uploaded image and generates new content that matches the style, colours, and similar elements present in the reference.
Instead of meticulously crafting a detailed prompt description including the mood, lighting, and colour scheme of the image, users can simply provide an example image, and Photoshop will use it as a visual guide to creating a new image with comparable characteristics.
Another major feature is auto-stylization powered by a custom-style engine. This allows Firefly 3 to automatically apply a wide variety of styles, colours, backgrounds and poses to your AI-generated imagery. It gives creatives much more control and personalization over the look and feel.
Combining these two features gives the user complete artistic control to modify and personalize all creations without worrying about generating appropriate prompts or matching the creative tone of the user.
Firefly Image 3 also has a better understanding of text prompts and scenes, enabling image generations that accurately reflect long, complex prompts and include richer details. Text renderings are also improved, resulting in image generations with clear text displays. This makes ideation for assets like posters more efficient.
Conclusion
Improved quality over the previous versions was the goal of this release and the consensus is that Image 3 has a wide range of applications when combined with the rest of the Firefly AI suite. While better image generation models exist, the combination of powerful Adobe tools with integrated AI features is highly valuable!
