Exploring new Midjourney zoom out feature

Introducing Midjourney V5.2

The eagerly awaited update that does not disappoint. Although Midjourney excels at accurate close-ups, the intricate backgrounds made full-body characters difficult to render. The most recent release, however, introduces significant improvements that are changing AI image generation and pushing the limits of creativity.

With the latest version of Midjourney, users are introduced to an exciting AI functionality called "Zoom Out." This feature goes beyond simply enlarging an image; it empowers users to expand the perspective and scope of their creations. By zooming out, users can create captivating scenes that extend beyond the original image, immersing viewers in a larger context.

In today's exploration of Midjourney's "Zoom Out" feature, we will put it to the test with advanced techniques that unlock its full potential. By combining custom prompts with this powerful functionality, we can craft engaging narratives that unfold across multiple scenes.

Zoom Out

I kept the default square aspect ratio to preserve the way the AI generates the image. Changing the aspect ratio can have a significant impact on the composition and overall appearance of the generated image. I wanted to explore the AI's capabilities within the constraints of the original aspect ratio to see how it would interpret and translate the prompt.

Prompt: raw photo of a beautiful woman

Midjourney V5.2 generates high-quality photos effortlessly with very simple prompts, showcasing advancements in AI technology for stunning, detailed, and visually appealing results.

Zoom out 1.5x

The AI's remarkable consistency is evident as it reproduces the same photo while skillfully incorporating a new background, showcasing its ability to create cohesive and visually appealing compositions.

Zoom out 2x

The AI's as it successfully generates a high-quality image while preserving the integrity and essence of the original. The ability to maintain the original image's quality demonstrates the AI's proficiency in producing impressive results.

Note that Zoom Out does not increase the maximum 1024px x 1024 px size of an image.

Upon closer observation, it appears that the images we have zoomed out 2x still can be zoomed-out. This prompts us to further challenge the AI's capabilities by pushing it to its maximum limits. Let's explore the extent of what the AI can achieve.

Zoom out 4x

While the overall quality of the photo remains impressive, there is a noticeable discrepancy in the length of her arm, which appears to be excessively elongated. This highlights a limitation in the AI's accuracy when it comes to anatomical proportions.

Zoom out 8x

The AI's performance has significantly deviated from the intended outcome, resulting in nonsensical and distorted images that no longer align with the original context.

Let's step it up a notch and explore a seriously challenging prompt. We may test the AI's capacity to create complex and sophisticated graphics by presenting it with tough, multifaceted inputs. It's time to test the limits and see how the AI can produce truly outstanding and inventive art. 

Prompt: a close-up shot of a samurai robot holding a sword on a cyberpunk city background, in the style of mecha anime, dynamic pose, intricate details, detailed background, 8k UHD --niji 5 --style scenic

Prompt Breakdown:

1- Samurai robot, mecha: Because of their intricate armor and complex shapes, samurai mechas present a significant challenge to the AI.

2- Holding a sword: Hands and swords have proven to be challenging for AI art generation. The intricate details, proportions, and complex shapes of hands and swords make them difficult to replicate realistically.

3- Cyberpunk city background: Cyberpunk cities are well known for their colorful, futuristic environments. These futuristic metropolises, dominated by soaring skyscrapers, feature a stunning assortment of neon lights, holographic displays, and cutting-edge technology.

4- Intricate Details: The prompt's inclusion of intricate details will force the AI to generate the image with increasingly complex details.

5- Niji 5 and Style Scenic: My favorite and most used AI, Niji, was also put to the test because I couldn't help myself. The Scenic style is known for its capacity to generate incredibly detailed and captivating environments.

Generating a quality image was quite challenging. The complexity of this prompt made the task difficult for the AI, showing that intricate prompts can be a challenge.

Zoom out 2x

 The generated image is disappointing. The sword's scabbard is excessively long, and the people on the street appear to be of poor quality.

                                 

Other v5.2 Features

make square:

Make Square provides you with the ability to modify the aspect ratio of non-square images, transforming them into perfect squares. For wide (landscape) images, the tool expands them vertically, while for tall (portrait) images, it expands them horizontally

You can easily identify the direction of expansion through the ↔️ and ↕️ emojis displayed next to the Make Square button. This convenient feature ensures that your images maintain a balanced and aesthetically pleasing composition. Look for the ↔️ Make Square button below any non-square upscaled images to access this functionality.

Custom Zoom:

The 🔎 Custom Zoom button empowers you to control the degree of zooming out on an image. Simply click the button located below an upscaled image, and a dialogue box will appear, allowing you to enter a custom value for the --zoom parameter. You can specify a value between 1 and 2, tailoring the zoom level to your specific requirements. This feature grants you precise control over the extent of zoom applied to your images, enabling you to achieve the desired visual effect.