The era of AI images has begun in earnest, and there are plenty of indications that we will be using this more and more. In addition to tools such as Dall-E 2 and Midjourney gaining prominence, a pretty interesting player in the market turns out to be Adobe’s new product Adobe Firefly. As a proud beta tester, I am sharing my first-hand opinion. I invite you to review this tool!
AI graphics generators are the future that will not be avoided
The boom in AI-enabled solutions is growing steadily, and the visual and graphics segment has not been spared. In addition to the solutions already gaining ground, Adobe Firefly is particularly worth watching. Why? I think it stands out for its really high quality and human-like appearance. First of all, the power of the solution from Adobe draws on legitimate sources and at the same time its own resources from Adobe Stock. What does this mean in practice? Other AI graphics generation software is based on open-source content available on the Internet. The solution from Adobe is based on their own legal resources to which they have copyrights.
Types of AI graphics generated
The following types of categories are available in Adobe Firefly:
- text to image
- generative fill
- text effects
- generative recolor
- 3D to image
- extend image
- personalized results
- text to vector
- text to pattern
- text to brush
- sketch to image
- text to template
Adobe Firefly in beta
The tool is in beta, but you can use almost all available features and its power.
How to write a prompt to generate AI graphics via Adobe Firefly?
Creating commands to generate AI graphics is called a prompt. How do you create them?
- To create a good referral, first, be specific. Specify exactly what you need, e.g. a dancing ballerina.
- Add all the characteristics you want to include in your graphic, e.g. a ballerina in a pink dress dancing in front of an audience.
- Specify the background and side elements of the graphic, the more detailed the better, e.g. a ballerina dancing by a lake in front of a large audience.
- Add an emotion or expression that is more “human”, e.g. an emotional audience, a happy ballerina, or an ethereal wavy skirt.
- Specify the style of the graphic, e.g. photo, portrait, landscape, landscape, graphic, animation.
Example prompt:
Example results:
The effects of some are slightly… taken out of a horror movie. The main reason is, first of all, AI’s still difficult to understand of the whole context and rely on what is publicly available in Adobe Stock. And these, as you know, can be very different, although still some of the best of all available image banks.
How about your experiences in testing AI in everyday tasks?
PS. The photo for the graphic of this post was generated by AI.