Using Midjourney for Inspiration as an Artist and Compositor

Example of My Workflow

I rendered my first image in Midjourney July 10, 2023, over two and half years ago. As of today, I have generated over 25,500 images using this platform so I’m not really a novice when it comes to using this tool in my artistic workflow.

But the interest thing about my workflow with Midjourney is that I have never reviewed a YouTube or read an article anywhere on line that explains how to use the tool the way I use it. So, I thought I would begin to document some of my processes just in case other Artist wanted to learn the way I use and AI tool for my artistic workflow.

To be honest, I have generated “AI-assisted art” where I use the output of Midjourney 100% as the art. I have purposeful workflows for generating wall art like abstracts and backgrounds for my digital art workflows where the output from the prompt is the final result.

But that said, most of the time, I USE MY OUTPUT for INSPIRATION or ELEMENTS to be used in a composite. As a photographer and a person who competes nationally in photography image competitions, I need to “MAKE” my art, not generate the art.

I have to demonstrate the technical skill to composite several elements together to build a story. To bring the pieces together in a composition with perspective, tone, color, and lighting to tell that story. The elements or pieces, used in the artwork can come from a variety of sources; photos I’ve taken, stock photos, or AI-generated elements.

In addition to my image competitions, I also belong to several art groups. While there are no strict rules for entry like the print competition, I still enter composited work versus 100% AI-assisted art. Compositing is what I teach and I feel I should “walk the talk” with my public submissions.

For this article I’m going to outline the process I just went through to build a whimsical artwork for Mardi Gras. The basic process involves (1) getting a story prompt first, then (2) dissecting that story prompt into two piece; an environment or scene only prompt and a second character prompt. I also get some additional elements that can be rendered that would fit the scene.

Here is the discussion I had with CHATgpt. Part of this dialog today was my retraining of the bot to the newest format I want to use for prompting based on research and articles I had read about the latest model and the best way to prompt V7 Midjourney.

I have learned some new information about the latest model in Midjourney and about how it expects prompts to be written. I learned that the newest model needs to have the "intention" of the image in the first sentence of the prompt. That first sentence should include the subject, the action or what is happening in the image and role of the scene or where the subject is. Then the prompt can include supporting information with descriptive information about lighting, mood, clothing, or anything else. But Midjorney will know "Who" "What" and "Where" from this first sentence.

I asked the bot if it understood what I was asking and it confirmed with a summary of what I had identified as the new prompt format I wanted to use. I then asked it for my story prompts.

Use this format for the STORY. Then under each story, break the story into two sub prompts, one for background where you only use the words that described the environment or scene. And another sub prompt for the description of the character. Then I can render the story, the scene and the character as three different jobs in Midjourney. Finally, add a short list of additional elements that might be in the scene that I can render on a white background to supplement the scene. Please build 5 whimsical Mardi Gras stories.

Here is one example of the 5 Stories that were returned and an example of one of my images from Midjourney:

The Peacock of the Parade

STORY PROMPT

A flamboyant anthropomorphic peacock leads a Mardi Gras parade down a vibrant New Orleans street at golden hour.

Sunlight filtering between historic buildings, confetti swirling in warm air, decorative balconies, purple and emerald banners draped overhead, lively crowd suggested in soft blur, glowing late-afternoon light, rich jewel tones, dynamic movement, festive and theatrical atmosphere.

BACKGROUND PROMPT (Environment Only)

New Orleans street during Mardi Gras, historic buildings with wrought-iron balconies, purple and emerald banners overhead, golden hour sunlight, swirling confetti, festive street decorations, warm glowing atmosphere, soft crowd blur in distance.

CHARACTER PROMPT (Subject Only)

Anthropomorphic peacock with dramatic fanned tail feathers in purple, green, and gold, wearing an ornate parade marshal jacket with gold braiding, tall feathered hat, holding a decorative baton, confident walking pose, theatrical elegance, layered feather textures.

Additional White-Background Elements

Decorative parade baton
Mardi Gras banner strip
Confetti cluster
Feathered shoulder epaulettes
Bead necklace bundle

I then built the Scene and the Background using a couple of my moodboards (Whimsical and Simple Surreal) and I rendered the Characters and the Elements using two different moodboards ( White Whimsy Isolate and Surreal Whimsy). This time I also searched for a whimsical Mardi gras style reference and added that code into the character build to get some great Mardi Gras colors to the characters clothing.

I upscale my images with Topaz Gigapixel and cut all my isolation off their backgrounds with Photoshop.

When I’m ready to build the artwork, I will select the background I like from those rendered, select my character(s) and elements and move on to a composite layout, toning, color matching etc in Photoshop and complete the artwork.

Categories: : Midjourney