Battle of the Prompts: Midjourney V5 vs V3

Table of Contents

1. Introduction
2. The Problem with Version 5 of Mid-Journey
3. The Importance of Legibility in Prompts
4. Testing Different Prompts in Version 5
5. Using Chat GPT to Improve Prompts
6. The Challenge of Recreating the Original Prompt in Version 5
7. Experimenting with Different Prompt Variations
8. Achieving Success with Pancake Skull Renders
9. Conclusion

Introduction

In this article, we will explore the process of creating realistic pancake skull renders using AI art technology and the challenges encountered along the way. We will specifically focus on the differences between version 3 and version 5 of mid-journey, a popular AI model for generating images Based on prompts. The goal is to recreate the original prompt from version 3 in version 5 and evaluate the results. Let's dive in and uncover the pros and cons of using different prompts in AI art.

The Problem with Version 5 of Mid-Journey

When the author first attempted to recreate the pancake skull render using version 5 of mid-journey, they encountered unexpected challenges. The initial results failed to capture the essence of the prompt, with the AI Generating images that resembled blueberry skulls instead of blueberry pancakes. Despite tweaking the prompts and parameters, the desired outcome was not achieved, leaving the author surprised and frustrated. This setback led them to question the effectiveness of version 5 compared to version 3.

The Importance of Legibility in Prompts

One crucial aspect of creating AI-generated art is ensuring that the images are legible and Instantly recognizable to the viewer. The original prompt for the pancake skull render aimed to produce a photorealistic image of blueberry pancakes in the Shape of a skull. This prompt needed to convey the concept clearly, allowing viewers to identify the subject matter at a glance. Legibility is essential in AI art, as the artwork should be understandable and compelling to a wide audience, potentially including clients or customers.

Testing Different Prompts in Version 5

To overcome the challenges faced with version 5 of mid-journey, the author experimented with various prompts to find a successful combination. They tried modifying the prompt by adding quotes, removing specific elements, or adjusting parameters to enhance or diminish certain features. This iterative process allowed them to explore different possibilities and understand how version 5 responded to different inputs. However, despite several attempts, the desired pancake skull render continued to elude them.

Using Chat GPT to Improve Prompts

Seeking an alternative solution, the author turned to Chat GPT, another AI model designed for generating text-based responses. They utilized Chat GPT to Create descriptive prompts for the pancake skull render, hoping to Elicit better results. By feeding in prompts that specified the desired characteristics of the image, such as stack of pancakes arranged in the shape of a skull, they aimed to obtain detailed and accurate outputs. The collaboration between mid-journey and Chat GPT showcased the potential for combining different AI models to improve prompt-based outputs.

The Challenge of Recreating the Original Prompt in Version 5

Recreating the original pancake skull prompt from version 3 in version 5 proved to be an arduous task. The author spent hours tweaking and refining the prompts, adjusting parameters, and iterating on the process to achieve satisfying results. They faced difficulties in balancing the legibility of the pancakes, the texture of the skull, and the placement of blueberries and maple syrup. However, they persisted in their efforts, determined to recreate the original intent and maintain the quality of the render.

Experimenting with Different Prompt Variations

Throughout the process, the author explored various prompt variations, ranging from close-up shots to macro photography, emphasizing different aspects of the pancake skull render. They adjusted parameters such as the presence of blueberries, the roundness of the pancakes, and the level of toasting to find the optimal combination. Despite encountering ups and downs along the way, they discovered several renders that showcased promise and demonstrated improved legibility as a pancake skull.

Achieving Success with Pancake Skull Renders

After numerous iterations and adjustments, the author eventually achieved success in recreating the pancake skull render in version 5. By incorporating all the essential elements – blueberries, maple syrup, and a well-defined skull shape – they produced compelling and visually appealing images. These renders managed to strike a balance between legibility and creativity, capturing the concept of a pancake skull while maintaining the photorealistic quality expected in AI-generated art.

Conclusion

The process of recreating the pancake skull render in version 5 of mid-journey proved to be a challenging yet rewarding experience. Despite the initial setbacks and frustrations, the author's dedication and perseverance led to the successful generation of photorealistic pancake skull renders that fulfilled the original intent. By experimenting with different prompts, adjusting parameters, and exploring the collaboration between mid-journey and Chat GPT, they demonstrated the potential for AI art to evolve and improve over time.