Using Midjourney for Landscape and Urban Design: A Practical Guide

Midjourney is changing how designers work. For landscape architects and urban designers, it offers powerful ways to visualize ideas quickly. This guide gives you a simple overview of how to get started and use Midjourney effectively to bring your design concepts to life.

Getting Set Up with Midjourney

Midjourney works through Discord, a free chat application. If you don't have Discord, you'll need to create an account and download the app or use the web version.

Once your Discord account is ready, go to the Midjourney website and click "Sign In." This will connect Midjourney to your Discord account. When prompted in Discord, authorize the application.

Creating Your Own Server

The main Midjourney server can be busy. To keep your work organized and focused, it's a good idea to create your own private Discord server and add the Midjourney bot to it. Here’s how:

• In Discord, click the plus (+) button on the left sidebar.

• Select "Create My Own" and choose your server type (or skip this step).

• Give your server a name and click "Create."

• Now, go back to the main Midjourney Discord server.

• Click the "Show Member List" button in the top right corner.

• Find and select the "Midjourney Bot" in the member list.

• Click "Add App" and choose your new server from the list.

• Authorize the bot for your server.

The Midjourney bot should now appear in your server, ready for you to use privately.

Understanding Basic Midjourney Commands

Midjourney uses simple text commands in the message box. All commands start with a forward slash (/). To see available commands, just type `/` in the message box.

The Imagine Command: Making Images

The most common command is `/imagine`. This is how you turn text descriptions into images. After typing `/imagine`, you'll see a prompt box appear. Type what you want to see in the prompt.

For example: `/imagine prompt landscape architecture rendering, sunny day`

Midjourney will generate four different image options based on your text. This might take a minute or two.

Working with Your Generated Images

After Midjourney shows your four images, you'll see buttons below them. These buttons let you work with the images:

• U Buttons (U1, U2, U3, U4): "U" means Upscale. Clicking a U button for a specific image (1=top left, 2=top right, 3=bottom left, 4=bottom right) creates a larger, more detailed version of that single image. Once upscaled, you'll get more options like zooming out or panning.

• V Buttons (V1, V2, V3, V4): "V" means Variation. Clicking a V button for an image creates four new images that are similar in style and composition to the one you chose, but with slight differences.

• Refresh Button: This button looks like a refresh symbol. It reruns your original prompt to create a brand new set of four images.

To save a generated image, it's best to upscale it first, then click on the image in Discord to open it fully, and save it from there. You can also access your generated images on the Midjourney website under "My Images" for better quality downloads.

The Blend Command: Mixing Images

The `/blend` command lets you combine two to five images you provide into a single new image. Upload your images to Discord, then type `/blend`. You'll see boxes to drag or select your uploaded images. Add them and press enter.

The Describe Command: Learning from Images

Put an image into Discord, then type `/describe` and select the image. Midjourney will analyze the image and give you four text prompts that it thinks would create a similar image. This is helpful for understanding what words Midjourney uses and values when interpreting images.

Going Deeper: Crafting Your Imagine Prompts

Getting closer to your desired image requires understanding how to write more effective prompts. An `/imagine` prompt can be broken down into three main parts:

[OL]

[LI]Image URLs (Optional)[/LI]

[LI]Your Text Prompt[/LI]

[LI]Parameters (Optional)[/LI]

[/OL]

Using Image URLs

Starting your prompt with an image URL can help guide Midjourney. You can use a photo, a sketch, or a reference image. Midjourney will consider the style, composition, and colors of the image when generating based on your text prompt.

Example: `/imagine prompt [LINK url="http://url.to/your/image.jpg"] New York City street, natural daylight, brownstone buildings`

Including an image URL helps steer Midjourney towards a specific look or feel that's hard to achieve with text alone.

Writing Effective Text Prompts

• Length: A longer prompt can give you more control, but too many words can sometimes confuse the AI. Each word holds weight.

• Word Choice: Avoid words with multiple meanings (like "crane"). Be specific. Midjourney often handles singular objects and specific numbers better than plurals (e.g., "three trees" instead of "trees").

• Structure: You don't need perfect sentences. A list of descriptive words separated by commas works well. Short phrasing is often best.

• Word Order: The order of words is less critical than in older versions, but using hard breaks (`::`) can help. Two colons with a space tells Midjourney to treat the text before and after as separate ideas with equal importance.

To craft better prompts, think about the key characteristics you want in your image. For landscape and urban design, this might include:

• Subject (what's in the image?)

• Medium (rendering, photo, painting?)

• Composition (overhead view, street level?)

• Environment (city park, forest, plaza?)

• Lighting (sunny, cloudy, dusk?)

• Color Palette (earthy tones, vibrant colors?)

• Mood (calm, busy, dramatic?)

• Specific details (paving materials, furniture style, plant types)

Think about these elements and use descriptive words in your prompt. You can also use tools like ChatGPT to generate lists of descriptive words based on your needs.

Writing effective prompts takes practice. Experiment with different wording and see how Midjourney interprets it. Keep track of what works. This is where automating your workflow can save significant time and help you iterate faster. Consider exploring tools like the Midjourney Automation Suite from TitanXT to streamline your prompt testing and image generation process.

Understanding Parameters

Parameters are options added to the end of your prompt using double dashes (`--`). They modify how Midjourney generates the image.

Example: `/imagine prompt urban park rendering --aspect 16:9`

Here are some common and useful parameters:

• `--aspect` or `--ar [width:height]`: Sets the image ratio (e.g., `--ar 16:9` for widescreen, `--ar 1:1` is the default square).

• `--chaos [0-100]`: Changes how varied the four initial results are. Higher numbers give more unusual results (default is 0). Useful for early exploration.

• `--no [thing]`: Excludes something from the image (e.g., `--no cars, people`). Use this to explicitly tell Midjourney what *not* to include.

• `--quality [.25, .5, 1, 2]`: How much time Midjourney spends on the first result grid. Higher quality takes longer but can add more detail (default is 1).

• `--repeat [number]`: Reruns the same prompt multiple times (up to 4 for basic plans). Similar to clicking the refresh button multiple times.

• `--seed [number]`: Use a seed number from a previous image to generate new images with a similar look and composition. To get a seed number, react to a generated image with the envelope emoji (✉️).

• `--stop [10-100]`: Stops the generation process partway through. Useful for intentionally blurry or unfinished looks.

• `--style [raw]`: Changes the default Midjourney look. `--style raw` often results in a more photographic or less stylized image.

• `--stylize [0-1000]`: How much Midjourney applies its default aesthetic (default is 100). Lower values stick closer to your prompt; higher values allow more artistic interpretation.

• `--tile`: Creates an image that can be repeated like a seamless pattern. Great for textures.

• `--version` or `--v [1-6]`: Switches between different Midjourney model versions. Each version interprets prompts differently.

• `--iw [.5, 1, 2]`: Image Weight. Used with image URLs (`--iw`) controls how much influence the starting image has compared to your text prompt (default is 1. You can use .5, 1, or 2). Useful if the image is too strong or not strong enough.

• `--weird [0-1000]`: Creates unusual and strange images. Experiment to see the results.

Mastering parameters gives you fine-tuned control over your image output. Combining parameters with well-written prompts and image URLs is key to getting the images you envision.

Generating and refining images can be repetitive. Tools exist to help. Accelerate your design process by exploring the Midjourney Automation Suite from TitanXT. Automating parts of your Midjourney workflow can free up time for actual design thinking.

Practice and Exploration

Just like learning to sketch or use design software, becoming skilled with Midjourney takes practice. Spend time experimenting with different prompts, parameters, and methods. Pay attention to how specific words and combinations affect the results.

Think of it as training your eye and refining your language to communicate effectively with the AI. The more you practice, the better you'll get at predicting outcomes and achieving the visuals you need for your projects.

Conclusion

Midjourney is a powerful visualization tool for landscape architecture and urban design. By understanding how to navigate Discord, use the core commands like `/imagine`, `/blend`, and `/describe`, craft descriptive prompts, and utilize parameters, you can start creating images that reflect your design ideas.

Start with the basics, experiment often, and get familiar with how the AI interprets your inputs. With practice, Midjourney can become an important part of your design workflow, helping you explore and communicate concepts more effectively.

Ready to take your Midjourney skills further and explore automated ways to enhance your design workflow? Check out the Midjourney Automation Suite from TitanXT.