Text to Image in Photoshop is a new feature that allows users to translate descriptive text prompts into photorealistic images. It empowers users to easily spark and remix inspiration, bringing creative visions to life with just a few steps. This feature is part of the generative AI capabilities integrated into Photoshop, which also includes additional creative controls like Content Type, Style Effects, and Reference Image.

To use the improved Generative Fill feature in Photoshop, powered by the Adobe Firefly Image 3 Model, follow these steps:

1. Select the area you want to fill with new content.
2. Choose the Generative Fill option.
3. Use the precise controls to inform the output, such as prompt comprehension for complex descriptions and generation variety.
4. Optionally, add a Reference Image to guide the AI and refine the results.

This improved experience delivers better photographic image quality and allows for more intentional creation with higher accuracy, details, and realism.

The new generative AI features and controls in Photoshop include:

• Generate Similar: Allows users to create variations similar to a selected result that aligns with their vision.
• Generate Background: Seamlessly removes and replaces backgrounds with generated content that matches the subject's lighting, shadows, and perspective.
• Enhance Detail: Improves sharpness and preserves details in outputs generated by Generative Fill.

These features help users explore concepts, refine details, and customize their creations with greater direction.

The new generative AI features in Photoshop are available in the Photoshop (beta) app as of the announcement date mentioned in the Adobe Blog, which is April 23, 2024. Users can access these features by trying out the Photoshop beta app or updating their existing Photoshop app to the latest beta version.

The Reference Image feature works with Generative Fill in Photoshop by allowing users to use visual inspiration to inform the content of variations generated. When using Generative Fill or Generative Expand, you can add a Reference Image to help guide the AI. This provides more control over the generated content, helps reduce randomness, and eases the challenge of writing precise prompts that accurately translate your imagination into visual results.