AI image generation is a game changer for content creators. If you’re a freelancer, a designer or artist, or you’re simply creating visual content for your own business, there’s likely a good use case for bringing AI image generation into your content production workflow.
Spoiler: There are so many possibilities!
MidJourney is one of many AI image generators available right now that uses machine learning algorithms and AI models to create vibrant images based on text a user inputs with a text prompt (for MidJourney the command is simply named “/imagine“).
In this guide with videos, I break down AI image generation and outline my step-by-step process for using the tool MidJourney – and ultimately how I turn those images into imaginative animated assets.
Before we get into it, it’s worth talking about some of the applications of using an AI image generation tool. I won’t go too in depth (for that, go here), but here are just a few ways creators can use a tool like MidJourney to enhance creative process and visual branding:
- For brainstorming and ideation – Ever have a case of creator’s block? An AI image generator can give you some ideas to help spark creativity – or at the very least, show you what an imaginative concept might look like.
- For blogs and thumbnails – Content creators often struggle with finding the right image for a post, video, or article. With AI image generation, you can type exactly what you’re looking for as a text prompt to get a custom image for that content.
- For social content – Create custom images for your social channels – creative videos, carousels, animated content, and even ads.
- For marketing materials – Add custom generated visuals to things like brochures, sales decks, flyers, cards, and other types of business collateral.
- For video assets – Possibly my favorite application. Take the creative assets you generate and turn them into unique animations or video components. That’s what I’ll be diving into in this series.
For our overview of AI image generation (what it is, how it works), go here.
Now, let’s dive into the fun!
Episode 1: What does MidJourney do?
First things first: MidJourney is on Discord so you’ll need a Discord account to get started.
Once you join, you’ll find a few “newbie” channels on the MidJourney Discord server to get a feel for how people write text prompts. You can literally type in anything to generate any type of image.
The trick is being really specific about what it is you want it to make. Things like art style, aspect ratio and the amount of light you want featured is key. The only limitation of MidJourney is your imagination.
Episode 2: Using text prompts to generate AI images
The first video introduced MidJourney – and now we’re getting into text prompts (specifically how you type one in and create an image).
The text prompt for this video’s thumbnail is a dog wearing a baseball cap, holding a rag doll cat in the style of Roy Lichtenstein. Underneath the grid of images created, you’ll have options to refine your image.
RELATED STORY: AI generated content: The good, the bad the ugly
Episode 3: Clipping MidJourney images in Photoshop
Now that you have your image asset, it’s time to clip it and prepare it for animation. In this video, I show you how I bring the image into Photoshop and separate it from the background with a layer mask, giving us the image with an alpha channel.
Then, I use the curvature pen tool to start selecting around parts of the body I want to isolate, such as the head and tail of the dog. We’ll export all the different parts separately.
After I’ve exported each element of the dog and they’re saved right here in the pictures folder for animation And yes, you guessed it in the next clip I’ll be applying some simple movements to animate the dog.
Episode 4: Animating AI generated images in Moho
Once you’ve generated, refined, and removed the backgrounds from your AI generated images in photoshop – it’s time to start animating.
I like to bring my assets into Adobe AfterEffects to add a VFX overlay.
Episode 5: How to Write the Perfect AI Prompt
There are a few tricks to creating the perfect AI image generation prompt for a tool like MidJourney in Discord.
Here are the four things to include in your prompt to get the best image possible:
- The main subject
- Details and characteristics
- Lighting and colors
- Overall art style.
Here’s an example:
A grizzly bear walking in the woods, wearing a biker jacket, holding a Snickers bar, wearing glasses, a baseball cap would be the main subject. Starry dark sky shooting stars, hyper realism. 4K would be the desired details and characteristics. Volumetric lighting, purple glow would be the lighting and colors.MidJourney prompt
As you can see in the video, the image is missing a few details but the main points are there. By using the four fundamentals in your prompt, you’ll create a repeatable framework to get what you want.