Midjourney, the popular AI image generator, has released its highly anticipated 5.2 version, bringing a feature that has been long-requested by the community. Aptly named “Zoom Out,” this functionality allows users to expand on previously created images, a process commonly referred to as “outpainting” in other AI image generators.
With Midjourney’s “Zoom Out” feature, users can now visualise a broader context, revealing more of the surroundings within an image, such as the environment surrounding a portrait.
This latest update introduces two default zoom levels, 1.5x and 2x to cater to users’ preferences. For advanced users, Midjourney offers the flexibility to freely configure the zoom level in combination with the desired aspect ratio. Additionally, the introduction of the “Make Square” function enables the transformation of non-square images into perfectly square ones through the process of outpainting.
In addition to the “Zoom Out” feature, Midjourney v5.2 brings notable improvements in image quality and integrated prompt tips. The introduction of the “Aesthetics System” aims to enhance the overall sharpness and visual appeal of generated images.
Furthermore, the AI’s text understanding capabilities have been enhanced, resulting in image outputs that better align with the provided text prompts. Users will also notice that the “—stylize” command on Midjourney now exhibits a stronger effect comparable to the performance of version 3. The value range for this command has been set from 0 to 1000, granting users precise control over the stylisation process.
To amplify the creative possibilities, Midjourney has expanded the variety of image suggestions. While this enhancement offers users increased diversity, it may require running the same command multiple times to achieve the desired outcome. The introduction of the new “High Variation Mode,” enabled by default, generates a greater number of distinct image options for a single prompt, sparking more creative exploration.
Moreover, Midjourney v5.2 introduces the “/shorten” command for prompt analysis, empowering users to identify words within a prompt that are likely to have minimal or negligible impact on the generated output. This analysis enables users to optimise their prompts effectively, ensuring that each word contributes meaningfully to the image generation process.
The 5.2 version of Midjourney is currently available as a test version, with the feature enabled by default. Users who wish to enable or revert to a previous version can easily do so within the Discord platform using the “/settings” command.
This announcement from Midjourney closely follows the recent release of SDXL 0.9, the latest iteration of Stable Diffusion which is an open-source image generator. The continued advancements in AI image generation technology underscore the commitment to providing users with cutting-edge tools to unleash their creativity.
As Midjourney evolves, it continues to empower artists, designers, and enthusiasts alike, offering new possibilities and pushing the boundaries of AI-assisted creativity.