Serving technology enthusiasts for more than 25 years. TechSpot is a source of trusted tech advice and analysis.
The creative world of AI-powered video and image creation continues to evolve at an incredible pace. Adobe has unveiled a new version. Firefly models of image generation at the recent Adobe MAX in London. It was also not a surprise to learn that the company has built a new Firefly app (with iOS/Android mobile versions coming soon).
What surprised some people was the company’s decision to integrate image creation tools from other GenAI model makers, including Google and OpenAI, into the new Firefly application. Adobe explained that they heard from their customers loud and clear that while they appreciate the work Adobe put into the Firefly model’s ability to generate videos and images without copyright concerns, there are some situations where they want more options.
If a design company is working with a customer, the designers will likely want to use copyrighted logos, images, etc. from that client in order to create new material.
Because of the way Adobe’s Firefly models are trained and built – specifically without copyrighted material – that wouldn’t be possible to do with Firefly. Leveraging image creation models from other vendors, however, could potentially allow those image variations with copyrighted material to be created.
What Adobe didn’t mention, though, is that creative people like to work with as many variations as they can. Bottom line is, they want more options. Because each image generation model works a bit differently, the outputs they generate have their own unique characteristics. It only makes sense then that designers were likely using other models anyway to get access to those different “looks.”
But by making the other models an option within the Firefly app, Adobe retains a bit more control over the creative process. This is especially true now, because new hooks in the Firefly app allow designers to take whatever is created from there – regardless of the model used to create it – and seamlessly pass it over to other Adobe creative apps, such as Photoshop, Premiere, Illustrator, and others. Frankly, it’s a pretty clever way to turn what initially seemed like competitors into something akin to partners.
A related but important development that Adobe unveiled is a new capability it’s calling Firefly Boards-previously known as Project Concept-allows designers to create and share an unlimited-sized canvas of images, generated from any model or copied from any location in order to stimulate new ideas.
These images “moodboards” can be used in the Firefly application to iterate more ideas using AI based on the images placed on them. Adobe has found a way to get its feet wet in the image creation pool by using what could be copyrighted material. Adobe has also made it clear any images created in the Firefly application will have Content Credentials.
With these credentials, companies can verify whether the images created were created with Firefly models (which are copyright-free and safe for public use), or with models from partners (or non-AI sources). It’s another great way to allow companies to experiment while still providing them with the assurance that Adobe’s entire Firefly technology and strategy is built around commercial safety. Adobe announced that, in addition to the integrations with other models for image and video creation, it has also made several major improvements to its own models. Firefly Image Model 4 focuses on creating a more photorealistic output compared to its predecessors, and offers more control options than ever before. Still images can be created using a greater range of camera angles, zoom settings, and more.
To encourage more experimentation the company offered a new format to create images with a lower resolution of 540p, which can be produced much more quickly. Once designers have found something they like, then they can create a 1080p HD version for final use. The new Image Model 4 Ultra, a paid option with a higher resolution, is optimized for what Adobe calls “complex scenes with small structures.”
Both Firefly Video Models and Adobe’s Vector Models are now generally available with significant refinements to control options and resolution. All outputs can be created from a text prompt, just like other Firefly models. They are also commercially safe.
A text description of the output is one of Firefly’s most interesting and yet unnoticed features. It may seem simple, but it is very informative because it shows the model’s view of the output.
Or, to put it another way: it allows you to reverse engineer the text that created the image or video. This is a great tool to learn how Firefly works, as the exact wording in a text prompt is crucial to an image creation tool. “thinks.” You can also use it to train yourself to write better prompts when you enter them into Firefly.
Adobe has also released several APIs to allow enterprise application developers to tap into the capabilities provided by Firefly models. Text-to Image and Text-to Video APIs are exactly what they claim to be, allowing companies to create applications that quickly build libraries of creative material. The Avatar API was designed to allow companies to create explainer videos using a digital avatar as the narration.
Overall, it was an impressive set of announcements which extends the reach of Adobe Firefly’s tools into some new interesting areas. The company’s core value and differentiation is based on its ability to produce commercially safe content, which has been verified to be free from any copyright violations. This is a valuable tool for large organizations. At the same time, Adobe also recognized that the creative users in these organizations-Adobe’s core customers-need more flexibility in terms of how they work with AI-generated material. These announcements appear to be a step forward in addressing these issues, and they offer some interesting potential options for the near future.
Bob O’Donnell, the founder and chief analysts of TECHnalysis research, LLC is a technology consulting company that provides strategic consulting services and market research to the technology industry as well as professional financial community. You can follow him at X @bobodtech
