Get Better Results: Understanding Midjourney Multi-Prompts and Weights

A Midjourney generated image using Midjourney Automation Suite

Creating images with text prompts in Midjourney is powerful. But how Midjourney reads your words makes a big difference. Simply typing a few words together doesn't tell the AI which ideas are most important.

Let's look at how you can guide Midjourney more precisely to get the images you want.

Simple Prompts: One Concept

When you type words like "still life painting", Midjourney sees this generally as one main idea: a still life painting. Adding commas or extra spaces between the words usually doesn't change this. The AI treats the whole phrase as a single concept.

Multi-Prompts: Separating Ideas

To tell Midjourney to see individual ideas within your prompt, you use double colons `::`. For example, "still:: life:: painting::" fundamentally changes how Midjourney interprets the prompt. Instead of one idea, it now sees three distinct concepts: "still", "life", and "painting". Midjourney then blends these separate ideas together in the final image.

Think about "sun flower". Without `::`, it's usually seen as one concept, giving you a traditional sunflower image. But "sun:: flower::" separates them into two concepts, leading to a different image that mixes the ideas of the sun and a flower in a less literal way.

Using Weights: Giving Concepts Importance

You can tell Midjourney which concepts are more important by adding numbers after the double colons. This is called using prompt weights. The numbers are relative to each other, not absolute values. For example, in "sun::1 flower::5", Midjourney will put much more emphasis on the "flower" concept than the "sun" concept because 5 is much larger than 1.

If you tried "sun::5 flower::1", the image would show more focus on the sun concept. This allows you to fine-tune the blend of your ideas. A weight of 1 is the default if you don't add a number. So "car:: blue::" is the same as "car::1 blue::1". The ratio matters: 2::1 is the same relationship as 4::2 or 100::50.

Understanding prompt weights lets you control the influence of each part of your prompt. If you want a blue car, using "car blue" as a single concept is often the direct way. Using "car:: blue::" treats them as separate ideas that are mixed, which might not always yield a straightforward blue car unless weighted carefully.

Tools that automate Midjourney prompting can help you experiment with different weights and multi-prompt structures more easily. Consider exploring options like the Midjourney Automation Suite from TitanXT to streamline your workflow.

Negative Weights: Less of Something

Sometimes you want to actively reduce the presence of a concept. You can do this with negative weights, using a minus sign before the weight number. For instance, "still life painting:: green::-0.5" tells Midjourney to decrease the amount of the color "green" in the image.

Negative weights work similarly to the `--no` parameter in Midjourney. They won't always completely remove the concept, but they will significantly lessen its influence on the final image.

Putting It Together

Instead of relying on commas or simple spacing, remember these key points for more control over your Midjourney images:

• Use `::` to separate different concepts you want Midjourney to blend.

• Use `::number` after a concept to give it more or less importance. These weights are relative ratios.

• Use `::-number` to reduce the presence of a specific concept.

Experimenting with multi-prompts and weights can lead to surprising and unique results. It's a powerful way to move beyond simple descriptions and start directing the AI more effectively.

Ready to take your Midjourney prompting to the next level? Automating your prompt variations can save you time and help you discover the best combinations. Check out the Midjourney Automation Suite to see how it can assist your creative process.

Conclusion

Mastering Midjourney doesn't require complex language, but rather understanding how the AI processes your input. By using double colons and prompt weights, you gain finer control over the concepts in your images, leading to more predictable and desired outcomes.