Midjourney Parameters and Settings for Beginners and Advanced Users

In this newsletter, read about:

• 🕵️‍♀️ Midjourney Parameters and Settings

• 🗞 News and Top Reads

• 📌 AI Art Tutorial: Mastering ControlNet on Stable DIffusion

• 🎨 Featured Artist: Chris Maestas

• 🖼 AI-Assisted Artwork of the Week

• 🤓 How to Get Started with Generative AI?

🕵️‍♀️ An Overview of Midjourney Parameters and Settings

These days, Midjourney can only be accessed via Discord, which significantly limits the tool’s ability to introduce a user-friendly interface. Still, there is a variety of parameters that you can play with to get better control over the output images. Unfortunately, most of the users are not aware of this rich toolkit they have access to.

In this post, I want to briefly overview Midjourney parameters and settings, and show how to be more efficient with your image generation by setting and reusing preferred parameters and specific prompt parts.

Settings

Several key Midjourney parameters can be accessed via the /settings command. Type the command, press Enter, and you’ll see the following options.

As you see, with these settings menu, you can:

• choose a model between v1, v2, v3, v4, v5, v5.1 and anime-focused Niji v4 or Niji v5

  • activate Raw mode for v5.1

• choose how much of the Midjourney’s artistic style you want to be applied to your images (stylize low, stylize med, stylize high, or stylize very high)

• switch between Public mode and Private mode, if your membership plan allows this

• switch between Fast mode and Relax mode, if your membership plan allows this

• switch on Remix mode, if you want to modify your prompts for variations and re-rolls.

The settings you choose will be applied to all your subsequent generations by default.

Beyond these settings, there are many more parameters that you can access by adding certain options to your text prompt. Let’s review the most important ones that work with the latest Midjourney versions 5 and 5.1.

Parameters

You can see the entire parameters list in the Midjourney documentation. Additional explanations are also available in the #prompt-faqs channel.

Here I’ll summarize basic information while focusing on version 5/5.1. Note that in Midjourney, some parameters work differently for different versions, or are available for only early/late versions. So if you are experimenting with earlier versions, make sure to check the documentation and FAQs channel for more relevant information.

When applicable, I’ll provide the default value of each parameter, possible range, and whether it works only for initial image generation or also for the subsequent variations and re-rolls.

Aspect ratio

The aspect ratio is the width-to-height ratio of an image. To set an aspect ratio for your image, type --ar or --aspect, followed by space and two numbers (width and height) separated by a colon, e.g. 4:5, 16:9, 1:1.

• You can use only whole numbers for this parameter. So for example, if you need 1.45:1 ratio, use 145:100 instead.

• Popular choices of aspect ratios include 1:1 for squared images, 3:2 for horizontal images, 2:3 for vertical images. Also, for social media and article illustrations, people often use 4:5 and 16:9 ratios.

• By default, Midjourney generates squared images with 1:1 aspect ratio.

• The aspect ratio that you specify in your prompt, will be also applied to all the subsequent re-rolls and variations you request.

Quality

The --quality or --q parameter defines how much time is spent generating an image, and hence, how much GPU time is spent per job. While architectural and some other images often benefit from the details of high-quality images, abstract images can sometimes look more interesting with lower quality.

• The quality parameter doesn’t impact resolution.

• The default value is 1.

• The other acceptable values are 0.25 and 0.5.

• This parameter only influences the initial image generation and is not applied to the subsequent re-rolls and variations.

Stylize

The --stylize or --s parameter defines how much of the Midjourney’s artistic style you want to be applied to your images. We have already mentioned this parameter in the settings section, where you learned that you can choose between four options, from style low to style very high. But in fact, you have even more flexibility as it’s possible to set the stylization level as a number ranging from 0 to 1000.

• With lower stylization values, the output images will be a better match to the prompt but less artistic. With higher stylization values, images will be very artistic, but less connected to the prompts.

• The default level is 100 (low stylization).

• Note, that when added to the prompt, this parameter effects only the initial generation and not the subsequent variations and re-rolls.

Chaos

The --chaos parameter defines how varied will be your output images.

• The default value is 0.

• The range of possible values is from 0 to 100.

• The higher the value, the more unusual and unexpected results and compositions you’ll get.

• Lower values produce more reliable and repeatable results.

• The --chaos parameter impacts only the initial image generation.

No

With the --no parameter, you can tell the Midjourney Bot what not to include in your image. If instead of using this parameter, you’ll just ask in the text prompt to exclude something, it may in fact lead to the opposite results because Midjourney seeks to depict everything it finds in the prompt.

• This parameter accepts multiple words separated by commas.

• It is applied to the initial image generation and all the subsequent re-rolls and variations.

Stop

You can use the --stop parameter if you want to finish the job partway through the process. Jobs stopped at an earlier percentage will create blurrier, less detailed results.

• The default value for the --stop parameter is 100 (full completion).

• The range of possible values is from 10 to 100.

• This parameter only works for the initial image generation and doesn’t work for upscaling.

Style

The --style parameter alters the aesthetic of output images, but it is only applicable to v5.1 and Niji v5.

With v 5.1 you can use --style raw to reduce the impact of the Midjourney aesthetic and get more photographic images.

With Niji v5, you can use:

• --style cute to create more charming characters and settings;

• --style expressive to get more sophisticated illustrations;

• --style scenic for more interesting backgrounds and cinematic character moments;

• --style original to use the original Niji v5, which was the default before May 26, 2023.

Note that this parameter has an impact on the initial generation as well as on all the subsequent variations and re-rolls.

Tile

The --tile parameter is used to generate images with repetitive patterns. This can be useful for creating background textures, wallpapers, fabrics, etc.

• Note that Midjourney generates only a single tile, and if you want to see the tile repeat, you’ll need to use a pattern-making tool like for example, Seamless Pattern Checker.

• When you add the --tile parameter to your prompt, it will be applied to your initial image generation and also to the subsequent variations and re-rolls.

Repeat

The --repeat or --r parameter simply runs your job multiple times.

• The range of possible values for Standard subscribers is from 2 to 10 and for Pro subscribers – from 2 to 40.

• This parameter can be only used in Fast GPU mode.

• The --repeat parameter is applied only to the initial image generation, so if you request a subsequent re-roll, it will run your prompt only once.

Image weight

When you use a reference image as part of your prompt, you can leverage the image weight parameter --iw to specify the relative importance of the image part of your prompt.

• The default value is 1.

• The range of accepted values is from 0 to 2.

• The higher the value of the --iw parameter, the higher would be the impact of the image on your final result.

• This parameter affects only the initial image generation, not the subsequent re-rolls.

Seed

When generating each image grid, the AI algorithm behind Midjourney uses a specific seed number, a starting point to create an initial field of visual noise.

• When you use the same --seed parameter for the same prompt, you’ll get nearly identical results.

• If you use the same seed, but a different prompt, you’ll get somewhat altered results but still with some resemblance.

• You can pick up the seed yourself by specifying any whole number between 0 and 4294967295.

• Or you can use the seed from the image grid you liked.

  • To see the seed value for any image grid, react with the envelope emoji ✉️ to the corresponding image grid.

• Note that the --seed value only influences the initial image grid.

• Furthermore, seed numbers are not static and shouldn’t be relied upon between sessions.

Preferences

In this section, I want to briefly touch upon the opportunities you have to improve the efficiency of your prompting. If you have any specific preferences with regard to parameters or even textual parts of your prompts, you can save these preferences and re-use them when necessary.

Prefer suffix

If there’s a specific set of parameters that you want to be applied to all your image generations, you can create a preferred suffix.

To set a preferred suffix, type a command like this, but with the parameters of your choice:

Note that this command accepts only parameters.

If you want to clear a preferred suffix, go to /settings and click on Reset settings.

Prefer option set

Midjourney also allows you to create several preferred option sets that you can append to your prompts, only when appropriate.

For example, let’s say you often generate portraits of different subjects but with the same lighting conditions, aspect ratio, and stylization requirements. In this case, you can create the corresponding option set, name it “portrait” for example, and then just type --portrait at the end of each prompt instead of specifying all these requirements every time.

To set a new preferred option set:

1. Type the command /prefer option set.

2. Specify the name that you want to assign to this option set.

3. Click right next to the box, where you specified the name.

4. Type or click on value that will appear above the typing box.

5. Type all the text and parameters that you want to be included in this options set.

6. Press Enter.

To delete a custom option, use /prefer option set and leave the value field blank.

To use the corresponding option set, just add its name as a parameter to your prompt.

Wrap Up

I hope this guide will help you to enrich your experience with Midjourney as well as exercise more control over your image generations. The opportunities are endless. Happy prompting!

🗞 News and Top Reads

• Google introduced a new virtual try-on feature that will show you what clothes look like on a wide range of people and poses.

• Vimeo, a popular video hosting platform, says that its new AI tools will help marketers quickly create video announcements using text prompts.

  • Their AI script generation feature is powered by OpenAI’s language models.

  • The AI-powered text-based video editor that Vimeo is going to introduce next month will also identify filler words, long pauses, and awkward moments that users can choose to remove.

• Celebrities leverage generative AI to earn from advertising without actually participating in advertising campaigns.

  • Brands create digital replicas instead of working with the stars directly.

  • This allows for much higher flexibility and likely much lower budgets.

• NVIDIA CEO believes generative AI will “supercharge” creators across industries and content types.

  • The key benefit of generative AI is its ability to scale up content generation.

📌 AI Art Tutorial: Mastering ControlNet on Stable Diffusion

In this guide, Vladimir Chopine goes over ControlNet on Stable Diffusion and explains how to install it and how to use it to have real control over your image generations. This toolkit is not as beginner-friendly as Midjourney, but if you generate images for professional use, mastering ControlNet is one of the important steps in your professional development.

🎨 Featured Artist: Chris Maestas

Chris Maestas (0009) is an American artist and designer who has been interacting visually and physically with the urban environment of Los Angeles since the early 2000s. Chris’s new journey into AI-assisted art has impressed numerous art collectors from all over the world. See more of his work @_0_0_0_9_.

🖼 AI-Assisted Artwork of the Week

🤓 How to Get Started with AI Art?

1. DALL-E: Creating Images from Text – introduction to text-to-image generation.

2. The DALL-E 2 Prompt Book – a guidebook by OpenAI that explains how to effectively right prompts to generate images across different domains (e.g., photography, illustration, art history, 3D artwork).

3. All-In-One Guide For Midjourney: The Art Of Prompts – a comprehensive guide on creating effective Midjourney prompts by Karol Andruszków.

4. Best Midjourney Prompts – a guide that covers the basics of Midjourney prompts (e.g., which keywords to use to create abstract art, surreal art, minimalism, etc) as well as some more advanced options (e.g., keywords related to camera lenses and filters, imitating certain artists and photographers without using their names). Finally, they provide a list of 600+ creative text prompts for image generation.

5. Stable Diffusion Prompt Book – a prompt book prepared by OpenArt. The book discusses ideal prompt format, using modifiers to change the style, format, or perspective of the image, applying ”magic words” to improve image quality, adding negative prompts, and adjusting Stable Diffusion parameters.

Share Kiki and Mozart

If you like this newsletter and know somebody who might also like it, feel free to share this newsletter. Let’s have more people learn about AI art!

If you have been forwarded this email and you like it, please subscribe below. And welcome to the world of AI art!