How to Include Multiple Characters in Midjourney Scenes

Midjourney is great for creating amazing images, especially with features like Character Reference (CREF) to keep one person consistent. But what if you need two or more consistent characters interacting in the same picture? This can be tricky because CREF usually works on one character at a time.

Trying to use multiple CREF image links in a single prompt just blends them together. Thankfully, people have found clever ways around this. We will look at two main methods that let you add more than one familiar face to your Midjourney creations.

The Pan Method: Adding Characters Separately

One way to get multiple characters is by using the "Pan" feature. This method builds your scene character by character.

• Start with your first character created using CREF.

• Open the image in Discord and use one of the Pan buttons (left, right, up, or down).

• If you have Remix mode on in your Midjourney settings, a prompt box will pop up.

• In this new prompt box, describe the *entire* scene you want, including the *new* character you are adding in the direction you panned. Include the CREF image link for the *second* character.

For example, you could start with an astronaut image. Pan right, and then in the new prompt describe "two different astronauts, the astronaut on the right is in his mid-20s with a thin mustache, wearing a white space suit with a red and white visor." You would also include the CREF link for the second astronaut in this new prompt.

This method works and can put multiple characters into one expanded image. However, they are generated somewhat separately. This makes it hard to get them to interact closely, like embracing each other or holding hands.

VARY Pose Connect: A Better Way for Interaction

If you need your characters to actually pose together and interact, the VARY pose connect method is generally better. This technique involves starting with an initial layout and then using the VARY (Region) tool to refine different parts.

Managing these steps for multiple characters and variations can become time-consuming. The Midjourney Automation Suite from TitanXT is designed to help streamline repetitive tasks like these, making it easier to manage complex multi-character workflows.

Steps for VARY Pose Connect with Multiple Characters

• Create a starting image that shows your characters in separate sections, perhaps like a split-screen photo, showing different poses you want to combine or build from.

• Copy the URL of this starting image.

• Use the VARY button on the image and select the tool to paint over just one of the characters in the image.

• Paste the URL of your starting image into the beginning of the prompt line.

• Modify the rest of the prompt to describe the overall scene you want and how you want the *painted-over* character to appear or pose.

• Generate the variation. Now use the VARY button on the *new* image and paint over the *other* character.

• Again, paste the original starting image URL at the start of the prompt. Adjust the prompt to define the second character's pose or appearance in the scene.

• You might need to paint over sections or use prompts to insert a clean background or combine elements.

• Optionally, use the Zoom Out or Custom Zoom features to adjust the framing.

The idea here is that you use an initial image to guide the character's look and relative position, and then refine each character or section using VARY and prompt adjustments.

Trying different variations and managing all the image URLs and prompts can get complicated. Tools like the Midjourney Automation Suite from TitanXT can assist by helping keep track of images and automate prompt variations, potentially saving you a lot of manual effort.

Putting It Together

Let's look at an example using the VARY pose connect method. You could start with an image prompt like: "two different poses of an Asian-American couple, Jenny and Ken. They are embracing and smiling at each other, isolated on a blank white background, split-screen photo from two different angles."

Generate this initial pose image. Then, you would copy its URL. Use VARY to paint over Jenny, paste the URL, and adjust the prompt to refine Jenny's look or position in the final desired scene. Repeat for Ken by painting over him in the resulting image, pasting the starting URL, and adjusting the prompt for him.

By strategically using VARY and image prompts, you can gradually build a scene with multiple consistent characters in specific poses.

Conclusion

While Midjourney's direct CREF feature only handles one character, methods like Pan and VARY pose connect offer effective workarounds for including two or more consistent characters in your images. The Pan method suits scenes where characters are separate, while VARY pose connect is better for interaction.

Experiment with both techniques to see which works best for your needs. Remember that tools are available to help make the process smoother. Check out the Midjourney Automation Suite from TitanXT to explore how automation might simplify managing these multi-step workflows.