Midjourney: Iterate, Iterate...

When I first started using Midjourney, I knew I had an advantage. Years of experience in film, video, audio, image, and text creation meant I understood composition, lighting, and storytelling. But did that mean I was an expert in AI-generated imagery from day one? Absolutely not.

I wasn’t stepping into the unknown blindly—I had the skills to analyze and refine visuals—but learning to communicate with AI was a whole different challenge. What I had spent years perfecting through hands-on work with cameras, editing, and directing had to be translated into words—a structured, clear, and AI-friendly language.

It was a learning curve. And like any creative process, it took trial, error, and iteration.

From First Prompts to First Frustrations

At the beginning, I kept my prompts simple. A cyberpunk detective in a neon-lit city."

The results? Visually impressive, sure, but they lacked something. The detective’s face would sometimes be distorted, or their outfit didn’t match the vision in my mind. The AI was generating what I asked for, but not what I meant.

That’s when I realized that Midjourney isn’t just an image generator—it’s a tool that requires direction, much like working with a creative team on a film set. You can’t just say “make it look cool” and expect a masterpiece.

So, I adjusted my approach.

Instead of:"A futuristic warrior with glowing armor."

I refined it:"A cybernetic warrior wearing sleek, form-fitting battle armor with intricate glowing blue engravings, standing under heavy rain, neon reflections bouncing off the wet ground, cinematic lighting, 50mm close-up shot."

The difference was massive.

I was no longer just describing a character—I was directing the shot.

Iteration: The Secret to Precision

Even with my background, I didn’t get it right immediately. Midjourney is powerful, but like any creative tool, it requires refinement and patience.

• Some characters had mismatched features. I had to tweak my descriptions.

• Certain styles didn’t translate well. I experimented with different art references.

• Consistency was a challenge. I had to find ways to recreate the same character across different prompts.

  
  

Every mistake taught me something new. Just like in filmmaking, the first take is rarely the best one. It takes multiple iterations to get the perfect shot.

And with each attempt, I became more precise, more intentional, and more effective in guiding the AI.

Exploring Different Styles

One of the most exciting discoveries was seeing how Midjourney could interpret the same character in multiple styles.

I experimented with:

• Hyper-Realism: "A detailed close-up of a battle-scarred warrior, metallic armor reflecting the setting sun, deep wrinkles, cinematic depth of field."

• Hand-Drawn Sketch: "A black-and-white ink drawing of a rogue assassin, high contrast, cross-hatching shading, intricate details."

• Anime Style: "A neon-haired hacker in a cyberpunk world, vibrant anime colors, expressive large eyes, dynamic pose, Studio Ghibli-inspired."

• 3D Render: "A stylized Pixar-like knight, soft rounded edges, bright color palette, exaggerated proportions, dramatic lighting."

Each one required a different approach. Some styles needed looser, more interpretative prompts, while others demanded rigid precision.

The AI as a Creative Partner

At some point, I stopped seeing Midjourney as just a tool—it started feeling like a collaborator.

Sometimes, the AI surprised me. It added small details I hadn’t considered, or it presented a new composition that made the character feel more alive.

Other times, it didn’t quite get it right, and I had to guide it more carefully—just like working with a production team, where communication is everything.

But that’s what makes it exciting. It’s not about perfection from the start—it’s about learning, refining, and pushing creative boundaries.