Beginner’s Guide to Midjourney

Are you someone with a vibrant imagination yet lacking the technical skills to draw, color, or paint? Imagine being able to bring your ideas to life in a matter of seconds without the need for artistic expertise. Thanks to artificial intelligence (AI) image generators like Midjourney, incredible digital art can be created with just a text description of your envisioned image.

Curious about Midjourney? Wondering how it works and what it can be used for? This article aims to guide you through the usage of Midjourney, providing insights into its significance and the impact it holds for the future. Join me on a journey to explore the potential of this AI in unleashing your creativity.

Introduction to Generative AI

To gain a deep understanding of Midjourney, it is crucial to understand the realm of generative AI. This subset of artificial intelligence systems excels in producing varied content, spanning images, text, audio, code, and various data formats. These sophisticated systems are designed to analyze and incorporate patterns from existing data, utilizing this knowledge to generate new content of a similar nature.

Recent advancements have propelled certain intelligent computer systems to seamlessly comprehend and respond to natural language commands. Notable examples include well-known chatbots like ChatGPT and LLaMA, as well as artistic programs such as Stable Diffusion, Midjourney, and DALL-E. These programs have the capability to transform text into visual representations. The applications of generative AI are widespread, impacting industries such as art, writing, music, software development, product design, game design, healthcare, finance, marketing, and fashion. This versatility opens up expansive possibilities for fostering creativity and driving innovation.

History of Midjourney

Midjourney was conceived from the visionary endeavors of David Holz and his team at an independent research lab in San Francisco. Prior to establishing Midjourney, Holz managed a design business while concurrently pursuing studies in physics and mathematics at the University of North Carolina at Chapel Hill. Between 2009 and 2011, he immersed himself in a Ph.D. program in applied mathematics, actively contributing to projects at NASA and the Max Planck Institute. Choosing a different path from the Ph.D. program, Holz relocated to San Francisco and founded Leap Motion in 2011.

In August 2021, marking a decade since the inception of Leap Motion, Holz embarked on a new venture with the creation of Midjourney. The inaugural private demo of Midjourney was revealed in September 2021. Acknowledging the tool's immense potential, Holz made the decision to share it with the public in February 2022, presenting Midjourney as a Discord bot. This platform not only served as a collaborative space for the Midjourney community to explore their creativity but also extended an invitation to newcomers to join and contribute.

Fundamentally, Midjourney is driven by the aspiration to delve into unexplored realms of thought and enhance humanity's capacity for imagination.

What is Midjourney?

Midjourney stands out as a significant player in the current surge of machine learning-driven image generators, aligning itself with distinguished counterparts such as DALL-E and Stable Diffusion. Specializing as a text-to-image generative AI tool, Midjourney boasts meticulous engineering aimed at producing images based on user-provided textual prompts. In the field of AI, text-to-image generation involves creating images in response to textual descriptions. Midjourney's text-to-image generative AI model incorporates both a language model and a diffusion model, underscoring its advanced approach to this cutting-edge technology.

How Midjourney Works?

Midjourney leverages state-of-the-art machine learning technologies, particularly impactful language and diffusion models, to revolutionize the conversion of natural language descriptions into visually striking images. Let's delve into the mechanics of how this seamless integration unfolds:

Language Models

Language models are formidable networks that demonstrate exceptional prowess in comprehending and generating human language. These models draw upon a vast array of textual sources, encompassing everything from books and articles to various online content, to refine their linguistic capabilities. The language model aids Midjourney in understanding the subtle meanings within your input prompts by identifying essential keywords and themes. This information is then converted into a standardized vector or embedding, acting as a guide for the diffusion model to produce an image aligned with the contextual cues in your prompt.

Diffusion Models

The diffusion model represents a specialized category of generative models intricately designed to simulate a diffusion process, be it in the forward or reverse direction, specifically for image generation. Employing a series of probabilistic steps, diffusion models systematically convert random noise into a cohesive image. The generative process commences with an initial noise tensor, resembling a random array of values without any meaningful image representation. This idea is analogous to starting with a blank canvas and embellishing it with random splatters of paint, then seamlessly blending those splatters to introduce or eliminate colors.

Diffusion models replicate a meticulously controlled process where noise is systematically introduced or removed from the initial tensor over multiple steps. Through training, the model gains insights into how genuine images evolve with incremental noise addition. In the generation phase, this procedure is reversed, starting with noise and progressively eliminating it to approximate a real image.

When faced with a textual prompt like "a solar eclipse over a serene desert wasteland," the model begins by generating initial visual noise, reminiscent of the static on CRT televisions. Through a series of gradual steps employing latent diffusion, the model progressively removes the noise, eventually producing an image that reflects the specified scene. This intricate procedure usually takes a minute or two to reach its complete development. Abruptly halting the process before completion may lead to an image with persistent noise, lacking the necessary denoising steps.

What Midjourney Can Be Used For?

So far, you’ve learned about what Midjourney is and how it functions. Now, let’s go into what Midjourney can be used for. As you already know, Midjourney is a text-to-image generator that can create and modify images to your imagination’s desire. What you might not know is that you can use Midjourney’s tools for some real-life applications in some creative ways making it versatile for various careers and jobs. Here are some instances that Midjourney has been used for:

  1. Producing Photorealistic Images: Midjourney offers the capability to generate photorealistic images that closely resemble photographs captured by seasoned professionals. Through customizable parameters, Midjourney can replicate authentic lighting effects, shadows, reflections, colors, perspectives, and even the characteristics of the camera lens. This advanced image generation tool is advantageous for photojournalists and photographers, whether they specialize in weddings or commercial projects.

  2. Crafting Character Designs: Midjourney excels in producing diverse images portraying a character in various poses and expressions. While maintaining consistency may pose a challenge initially, it becomes achievable through dedicated time and practice in prompt creation. Professionals in the realms of concept art, illustration, and storyboarding within the gaming and film industries stand to derive significant advantages from the capabilities offered by Midjourney.

  3. Branding and Logo Design Services: Midjourney specializes in the creation of unique and creative logos aimed at enhancing brand visibility and effectively communicating brand messages. While Midjourney may not consistently deliver flawless and precise logos, it serves as an excellent tool for gathering ideas and inspiration. Professionals in fields such as graphic design, creative marketing, and web design can find valuable utility in utilizing Midjourney for their creative endeavors.

  4. Clothing Design: Midjourney serves as an exceptional tool for crafting distinctive patterns and designs suitable for T-Shirts and various apparel. Leveraging the capability to generate designs from either a description or a sketch, Midjourney AI has proven to be highly advantageous for fashion designers and professionals within the fashion industry.

  5. Visual Storytelling: Midjourney offers a valuable resource for artists seeking to elevate their storytelling through visual elements. It empowers creators to enhance narratives and create compelling illustrations for various mediums such as books, comics, manga, and graphic novels. With Midjourney, users can skillfully generate intricate illustrations of characters, settings, and key scenes by crafting specific prompts. This tool is particularly beneficial for novelists, screenwriters, and comic book creators, providing a versatile solution for enhancing visual storytelling in their respective fields.

  6. Interior Design Solutions: Midjourney offers a comprehensive approach to room and house renovations by creating visual representations aligned with furniture preferences, room layouts, and color schemes. It considers crucial elements such as lighting and space constraints, enabling users to experiment with diverse design options before making final decisions. Furniture designers and interior bloggers will discover Midjourney as a valuable asset in their creative pursuits.

  7. Architecture Design Solutions: Midjourney specializes in crafting architectural designs that seamlessly blend functionality with aesthetic appeal. It lies in generating innovative and inspiring ideas for buildings and structures, taking into account crucial factors such as space utilization, choice of materials, and environmental impact. Professionals across various domains, including architects, urban planners, and construction workers, have successfully integrated Midjourney AI into their workflows, experiencing the transformative power of our architectural solutions.

In light of various real-world situations, Midjourney emerges as a platform offering promising prospects for sustainable livelihoods. Despite its recent introduction, Midjourney boasts substantial untapped potential with versatile applications. The broad range of possibilities it presents unlocks numerous career opportunities spanning diverse creative fields. For individuals interested in harnessing the potential of Midjourney, the subsequent sections will offer guidance on the setup process.

Getting Started with Midjourney

In continuation of our previous conversation regarding the inception of Midjourney, it was initially presented to the public as a Discord bot. This bot will be discussed here in a little further in this article. Midjourney now operates a dedicated Discord server, allowing users to interact with the bot within the server or through direct messages. Should you be unfamiliar with Discord or have queries about bots and the exclusive accessibility of Midjourney through Discord, the Midjourney staff are available to offer clarification and address any questions you may have.

What is Discord?

Discord stands out as a widely utilized and free platform renowned for its efficacy in enabling online communication and collaboration. Functioning as a versatile hub for real-time text, video, and voice communication, Discord sets itself apart from other social platforms by structuring users into servers, each serving as a unique community. These servers offer the flexibility of being either public or private, empowering users to customize their experience according to their preferences.

Whether one is joining a large community centered around a shared interest or creating a smaller, private server for a select group of friends, Discord offers a high degree of flexibility in community building. Notably, the platform generates revenue by providing upgrades for user accounts or servers, eschewing the traditional model of ad-based monetization.

Widely embraced by gamers, streamers, content creators, and online learners, Discord's rich array of features and integrations significantly elevates the overall user experience. This versatility and functionality make Discord the preferred platform for those seeking interactive and engaging online interactions.

What are Bots?

Discord bots serve as advanced AI-driven tools, playing a pivotal role in automating various tasks within a Discord server. They significantly enhance community engagement and facilitate effective moderation by extending a welcoming gesture to new members, overseeing member interactions, and promptly addressing issues through the implementation of user bans for disruptive behavior. These adaptable AI systems are available in multiple languages, offering marketers and business managers the capability to automate a wide range of functions on the server. Moreover, Discord bots bring an entertaining element to the server environment, allowing users to integrate games, music, and other engaging content. This is exemplified by features like Midjourney.

Why Does Midjourney Use Discord?

Now, let’s go onto why Midjourney still uses Discord. There are several reasons why Midjourney runs through Discord.

The first rationale behind Midjourney's establishment is to create a central platform for both creators and enthusiasts. Functioning beyond mere utility, Midjourney serves as a nexus for cultivating creativity and self-expression. Users can actively participate in discussions on Discor, connecting with like-minded individuals who share a common interest in text-to-image AI. This platform facilitates the exchange of prompts and images, encourages constructive feedback, provides a space for seeking advice, and enables collaborative skill enhancement. Additionally, users can immerse themselves in weekly challenges, events, and contests organized by the Midjourney team, contributing to an enriched and dynamic experience.

The second rationale is to provide comprehensive support and guidance to users. As Midjourney is currently undergoing development and continual evolution, occasional bugs, errors, or issues may arise during image generation. However, users can take comfort in the fact that Midjourney has a dedicated team of staff and guides ready to assist in resolving any problems that may occur. The Midjourney staff highly values user feedback and welcomes proposals for new features, actively seeking ways to enhance both the bot and the overall service.

The third rationale is to ensure that users are well-informed through timely announcements and updates. In line with the dynamic nature of Midjourney's continuous evolution and improvement, users will receive real-time notifications regarding any changes, additions, or fixes to both the bot and the service. This approach allows users to gain sneak peeks into upcoming features and stay abreast of exciting plans for the future.

The ultimate reason is to offer a user-friendly and easily accessible experience in utilizing Midjourney and its bot. Unlike many AI generators that necessitate specific hardware or software requirements, Midjourney distinguishes itself by operating seamlessly through the Discord bot, eliminating the need for any specialized hardware or software installations. This simplicity enhances the accessibility and usability of the platform for a broader user base.

Setting Up Discord

Now that you understand what Discord is, how bots function, and why Midjourney integrates itself with Discord, it’s time to go through the process of setting up and getting started with Midjourney. To begin, you need to create a Discord account first. Follow the instructions below in order to create your Discord account:

  • Step one, go to: https://discord.com/download to download the app version or go to: https://discord.com/login to use directly from your internet browser. If you choose the internet browser option, you can skip ahead to step 5.

  • Step two, if you opted for the app version, you click the blue button that says “Download for Windows/Mac”. Refer to figure 1.1 below:

Figure 1.1 - Click “Download for Mac” if you’re on an Apple computer and “Download for Windows” if you’re on a Windows computer.

  • Step three, locate a file called DiscordSetup.exe in your downloads and double-click to start the installation process.

  • Step four, once installed, open the Discord app.

  • Step five, on the login page, click the blue “Register” link near the bottom below the blue “Log In” button. Refer to figure 1.2 below:

Figure 1.2 - Click the blue link labeled “Register” below the blue “Log In” button.

  • Step six, fill in the information for the account and then click “Continue”.

  • Step seven, you will need to claim your account by verifying your email address that you use to register. This is important to do in order to enjoy all chat functions Discord has to offer as well as ensuring you keep your username, discriminator(four digits next to your username), and remembering the servers you have joined. Check your inbox of the email you used to create your account. If you do not see an email, simply press the Resend button in the green banner at the top of the Discord app. Once you receive the email, click Verify Email and you’re good to go.

Setting Up Midjourney

  • Step one, visit this link: https://www.midjourney.com/account/

  • Step two, click "LogIn" and sign in with your Discord account.

  • Step three, authorize Midjourney Bot to access your Discord account when prompted in the pop-up window.

  • Step four, choose a subscription plan that suits your needs. Midjourney has four subscription tiers that are paid monthly or annually for a 20% discount. Each subscription plan includes some unique features and perks but all plans have access to the Midjourney member gallery, the official Discord, and general commercial usage terms. Please observe figure 2.1 below for plan comparisons:

Figure 2.1 - Midjourney’s current subscription plans.

  • Step five, once you have subscribed to a plan, return to the Discord app or browser window that has Discord open.

  • Step six, on the left-hand sidebar, the server list, press the green ‘+’ button. Refer to figure 2.2 below:

Figure 2.2 - Click Green ‘+’ button on the left-hand sidebar.

  • Step seven, a pop-up window will come up, click the “Join a Server” button. Refer to figure 2.3 below:

Figure 2.3 - Click on “Join A Server” button.

  • Step eight, paste or type the following URL: http://discord.gg/midjourney and press “Join Server”. Now, you’re ready to start creating your first images! Refer to figure 2.4 below:

Figure 2.4 - Paste or type http://discord.gg/midjourney and press “Join Server” button.

Interacting With Midjourney on Discord

Before you begin, it's essential to note that there are three distinct methods for interacting with the Midjourney bot to generate images.

The first approach includes accessing the Midjourney Discord server. Within the server interface, a left sidebar prominently showcases a list of text and voice channels. It is crucial to identify and select a channel labeled either "general-#" or "newbie-#." These specified channels cater to both beginners and seasoned users of the Midjourney bot, ensuring a smooth and enjoyable interaction experience. Observe figures 3.1 and 3.2 for “general-#” and “newbie-#.”

Figure 3.1 - Within the server, you’ll find the newbie rooms on the channels listed on the left.

Figure 3.2 - You’ll find the general rooms listed further down the channels list from the newbie rooms. You may not see these at first until you have begun generating your own images.

Engaging with the Midjourney bot through one-on-one direct messages represents the second method, offering a personalized and efficient interaction. To access this feature, navigate to the Midjourney Discord server, where you'll find an additional sidebar on the right, displaying a hierarchical list of users and bots. Although this list may initially be hidden, a simple click on the first button to the left of the search bar at the top right will reveal it. Once located in the list, right-click on the Midjourney bot, and select "Message" to initiate a direct communication. This method proves particularly advantageous for image generation, addressing the potential challenge of images getting lost in the general-# and newbie-# channels, where numerous contributions can clutter the space. While direct messages consolidate your generated images, it's essential to be aware that retrieving specific images may require navigating through potentially numerous direct message interactions. Refer to figure 3.3 below:

Figure 3.3 - Click the person icon at the top right by search bar to reveal the server’s hierarch of members. Right-click on Midjourney bot and click Message to set up a private direct message conversation with it.

Fortunately, there is a third method available for effectively utilizing the Midjourney bot. This approach enables you to systematically organize and manage your images in a structured manner. To engage with the Midjourney bot in a more focused way, free from any external contributions, consider creating a dedicated Discord server exclusively for yourself and the Midjourney bot as its sole members. By establishing such a server, you gain the ability to configure specific text channels, offering a platform to create and arrange images based on your preferences. Follow the outlined steps below to set up your personalized Discord server for optimal organization:

  • Step one, once again, on the left-hand sidebar, the server list, press the green ‘+’ button.

  • Step two, a pop-up window will come up, this time you will click the “Create My Own” button. Refer to figure 3.4 below:

Figure 3.4 - Once you have pressed the green ‘+’ button again, you will press “Create My Own” this time in order to create your server.

  • Step three, once you create the server, you should have a page like this for your server. Refer to figure 3.5 below:

Figure 3.5 - This is what your new server you created for yourself should look like. You should only have a general text channel and a general voice channel.

  • Step four, you need to add the bot to your server. To do this, go back to the Midjourney Discord server and find the Midjourney bot in the right sidebar again.

  • Step five, this time, click on the Midjourney bot and the user profile information should pop up.

  • Step six, near the top, click on “Add App”. Refer to figure 3.6 below:

Figure 3.6 - Find the Midjourney bot in the Midjourney Discord server again. Click on it then click gray “Add App” button.

  • Step seven, another window will pop up, select your server from the drop-down menu and hit “Continue”. Refer to figure 3.7 below:

Figure 3.7 - From the drop-down menu, select your new server you created and then click the blue “Continue”.

  • Step eight, now that you have Midjourney bot in your own server, you can add some more text channels and name them whatever you prefer to help you better organize by clicking + to the right of text channels. Refer to figure 3.8 below:

Figure 3.8 - Clicking the + to the right of text or voice channels will allow to add another text or voice channel to that category. Since it is just you and Midjourney bot, there should be no need for voice channels.

  • Step nine, set the name, channel type, and privacy for this channel then click “Create Channel”. Refer to figure 3.9 below:

Figure 3.9 - Leave channel type as text, give your text channel a name, and click the blue “Create Channel” button.

  • Step ten, repeat steps eight and nine to create more text channels for organizing your generated images.

Creating Your First Midjourney Images

As you embark on the creation of your initial images, effective interaction with the Midjourney bot on Discord is crucial. Utilize specific commands tailored to different purposes, including image generation, settings modification, and other valuable functions. Whether you're in the general# or newbie# channels within the Mid Journey Discord, engaging in direct one-on-one messaging with the Midjourney bot, or interacting via text channels in your private server, the process for image generation remains consistently applicable.

The universal command employed for image generation is /imagine. To utilize the /imagine command effectively, follow these steps:

  • Type ‘/imagine prompt:’ in the message field. You can also quickly select the /imagine command from a dropdown list of other commands once you type ‘/’.

  • Follow ‘/imagine prompt:’ by typing a description of the image in the prompt field.

  • Hit Enter to send your message.

Refer to image 4.1 below to see /imagine being used:

Image 4.1 - This is an example of how using /imagine should look like.

The automated system will carefully analyze the input you provide and commence the image generation process. This procedure harnesses advanced Graphics Processing Units (GPUs), with each image creation utilizing the allocated GPU time specified in your subscription plan. You can easily monitor the available and utilized GPU time by employing the /info command. Once your request is fulfilled, the Midjourney bot will promptly reply, presenting a grid containing four unique images generated in accordance with its interpretation of your input. Observe the generated images in figure 4.2 below:

Figure 4.2 - Executing the /imagine prompt, we were given four unique images generated by Midjourney.

Midjourney Tools

Observe figure 5.1 and then observe figure 5.2. Notice rows of buttons beneath your generated images that are labeled U1, U2, U3, U4, and V1, V2, V3, V4, and then a refresh button.

Figure 5.1 - Take note of how each image is numbered. This is important when taking in account Figure 5.2.

Figure 5.2 - The buttons in the red square are some of Midjourney’s Tools. The numbers behind each “U” or “V” represent the image referenced in Figure 5.1. If you select “U1”, you will get a single image of the top left image by itself.

In earlier iterations of Midjourney, the 'U' buttons were employed for the purpose of upscaling images. Presently, these 'U' buttons serve the function of facilitating image selection, allowing users to detach an image from the grid. This not only simplifies the downloading process but also provides enhanced editing and generation options. On the other hand, the 'V' buttons are designated for generating image variations, producing a new grid containing four distinct images that uphold the overall style and composition of the selected variation. The blue refresh button essentially re-executes the task, yielding another set of four images.

Now, imagine you've discovered an image you fancy and have isolated it from the others. However, you may want to make some enhancements or modifications to your image. Below this individual image, you'll find several options:

  • Vary(Strong): Creates a fresh grid of four images incorporating substantial and noteworthy alterations from the chosen image.

  • Vary(Subtle): Produces a fresh grid of four images with minimal adjustments from the selected image.

  • Vary(Regional): Enables precise selection of designated areas within the image for modification, while maintaining the integrity of the remaining sections. This feature is particularly advantageous, addressing instances where Midjourney may encounter challenges, especially in areas like hands and feet. Once your selections are made, this option will generate a new grid of four images focused on the specified regions requiring modification from the original image.

  • Upscale 2x and 4x: These functionalities enhance the dimensions of your image. It is important to note that the upscale process consumes GPU resources and may require a longer processing time to generate the enlarged image.

  • Zoom Out 1.5x, 2x, and Custom Zoom: These features broaden the visual scope of the image without altering its content. The expanded perspective is intelligently filled using cues from the original prompt and image. The 'Custom Zoom' option provides the flexibility to define a specific aspect ratio beyond the standard 1.5x or 2x options.

  • The arrows or pan buttons extend the image in a user-selected direction without altering the original content. The expanded pan area is intelligently populated based on cues from the original prompt and image.

  • The heart or favorite button serves as a tagging mechanism for identifying and retrieving your top-performing images conveniently through the Midjourney website.

  • The 'Web' button directs you to view your image on Midjourney.com.

Observe figure 5.3 below to see more Midjourney tools:

Figure 5.3 - Here are more of Midjourney tools you can utilize when working with a single image of your interest.

Additionally, you can use emoji reactions to trigger an action from the Midjourney bot:

  • The red 'X' reaction emoji functions to cancel or delete the job, removing it from the Midjourney website. This feature aids in managing clutter and eliminating undesired generated images.

  • The envelope reaction emoji facilitates the transmission of a completed job to direct messages. The direct message will include essential details such as the image’s seed number and Job ID. While the immediate significance may be minimal, this feature holds potential for future utilization in advanced prompt building. Refer to figure 5.4 below:

Figure 5.4 - Clicking the smiling face icon allows you access to emojis. There are only two that will trigger an action from the Midjourney bot.

Midjourney Commands

As previously mentioned, users can engage with the Midjourney bot on Discord by issuing commands. These commands serve as versatile tools, enabling users to interact with Midjourney in a manner aligned with their creative objectives and preferences. In addition to the /imagine command, there are various other commonly used commands and their respective purposes outlined below:

  • /ask: Offers responses to inquiries regarding the Midjourney. The Midjourney Discord platform features dedicated support channels aimed at addressing user questions and concerns.

  • /blend: Facilitates the upload of two to five images, subsequently integrating them seamlessly.

  • /describe: Generates four exemplary prompts derived from an uploaded image.

  • /fast: Enables the fast mode to expedite the process of image generation for enhanced efficiency.

  • /help: Presents informative foundational details and useful tips pertaining to the Midjourney bot.

  • /imagine: Produces an image based on a given prompt.

  • /info: Access details pertaining to your account and any associated tasks.

  • /private or /stealth: Provides the capability for your generated images to remain confidential on the Midjourney website. This exclusive feature is accessible to subscribers of the Pro or Mega plans.

  • /public: Facilitates the visibility of your generated images to other users on the Midjourney website.

  • /relax: Transitions to the relax mode to optimize image generation for a more deliberate and controlled pace.

  • /settings: Provides the capability to view and customize settings for the Midjourney bot according to your preferences.

  • /shorten: For lengthy prompts, receive recommendations on how to condense and streamline the content for enhanced conciseness.

  • /subscribe: Creates an individualized link directing users to their account page for subscription plan options.

  • /turbo: Transitions to turbo mode for accelerated image generation, surpassing the speed of the /fast mode. This advanced mode operates four times faster, albeit consuming twice the GPU minutes. Turbo mode is exclusively accessible in Midjourney version 5 and beyond.

Observe figure 6.1 below and notice when you type ‘/’ it will bring up the Midjourney Commands:

Figure 6.1 - When you type ‘/’ in the message box, it will bring a list of ‘/’ commands Midjourney uses.

Midjourney Parameters

In addition to commands, Midjourney offers customizable parameters that are integral in aligning the image generation process with your creative vision. These parameters empower you to meticulously adjust various facets of the generation, including aspect ratio, rendering quality, and style. Provided below is a compilation of commands that can be incorporated at the conclusion of your prompts to effectively manipulate the resulting image:

  • Aspect Ratio(--aspect or --ar <value>:<value>): This adjustment modifies the aspect ratio during the image generation process, ensuring that images are consistently produced at a 1:1 aspect ratio.

  • Image weight(--iw <0-2>): This parameter determines the relative importance of the image in comparison to the text prompt. Values exceeding 1 signify a higher priority on the image, whereas values below 1 prioritize the text prompt more. The default setting for image weight is 1.

  • No(--no): Referred to as negative prompting, this feature enables you to specify elements you wish to avoid in your images. Simply enumerate one or multiple items you want to exclude after the "--no" parameter, separating each item with a comma or space.

  • Quality(--quality or --q <.25, .5, or 1>): This parameter determines the duration allocated to the rendering process for image generation. A lower numerical value indicates reduced rendering time, albeit yielding a lower quality image. Conversely, a higher numerical value signifies an extended rendering time, leading to a higher quality image output.

  • Repeat(--repeat or --r <1-40>): This functionality allows for the creation of multiple tasks from a single prompt. Users have the option to specify the desired number of iterations for Midjourney to generate outputs based on the same prompt.

  • Stop(--stop <1-100>): This feature concludes the image generation process at the specified percentage set within your prompt. For instance, employing the "--stop 75" command instructs Midjourney to halt image generation upon reaching 75% completion. This can be effectively utilized to impart a Gaussian blur effect to your images.

  • Stylize(--stylize or --s <0-1000>): This parameter influences the extent to which Midjourney's training is applied to your images. Lower values result in images closely aligned with the prompt but may exhibit less artistic expression. Conversely, higher values yield highly artistic images, albeit with a reduced connection to the original prompt.

  • Niji(--niji or --style<original, cute, expressive, scenic>): This feature enables the utilization of an alternative model specifically dedicated to the generation of anime-style images.

  • Version(--version or --v <1, 2, 3, 4, 5, 5.1 or 5.2>): This functionality assists in the selection of a prior version of Midjourney, offering users the option to choose a preferred version over the current iteration.

Observe figure 7.1 below and notice how I added some parameters in my previous prompt:

Figure 7.1 - Notice I used a few parameters to help my original four generated images follow more closely to the original prompt and have a more anime style.

Midjourney Fundamentals

For newcomers to the platform, this wealth of information might appear overwhelming at first, possibly portraying Midjourney as a challenging undertaking. However, as you commence your journey, you'll gradually navigate through the complexities, investing time to create impressive images and explore innovative concepts. This section is designed to offer a comprehensive exploration of the fundamental aspects of crafting prompts for Midjourney.

Highlighting simplicity in your prompt descriptions is crucial. Begin with a clear and concise approach, progressively integrating complexity, additional details, and styles. It's important to note that Midjourney prioritizes the first 60 words of your prompt. Utilizing full sentences may lead to the exclusion of subsequent details. To address this, prioritize complete phrases or keywords. While not mandatory, the following offers a valuable guide for structuring your prompts:

[content type], [description + subject + adjectives], [style], [parameters]

Putting It All Together

Let's analyze an illustrative example below using the structured prompt format in figure 8.1 below:

Figure 8.1 - Red highlighted text represents the content type. Blue highlighted text represents the subject. Yellow highlighted text represents the theme and style. Green highlighted text represents the parameters used.

In the first part of the prompt, highlighted in red for clarity, key phrases like "full length," "character reference sheet," and "multiple dynamic action poses" effectively communicate the desired content type to Midjourney. Similar specifications can be applied for alternative content types, such as aerial perspectives, illustrations, or contemporary art. This ensures a clear and concise communication process.

The content type indicators highlighted in blue, descriptive elements such as "beautiful" and "strong female gunslinger," along with specifics like "long black hair," "smokey eyeshadow," and "revolvers," contribute to the portrayal of the subject. This segment is dedicated to expressing the subject's appearance, actions, or poses in a comprehensive manner.

Following this, the highlighted elements in yellow, including "western steampunk" and "in the Mamoru Oshii," clarify the intended theme and style. These aspects serve as guidance for Midjourney in creating images infused with distinct artistic influences, sometimes even mirroring the work of different artists.

Lastly, highlighted in green, are the parameters. As discussed earlier, these parameters, specifically represented as "--niji" for an anime art style and "--s 700" to empower Midjourney's artistic autonomy, play a significant role in refining different aspects of the image generation process. It is imperative to position these parameters at the conclusion of your prompts to guarantee their optimal application.

If you notice in that prompt above, I adhered to the foundational principles for formulating prompts within Midjourney. I maintained brevity by limiting the prompt to 60 words, employing concise phrases and keywords to guide Midjourney in generating the desired outcome. While this prompt structure is not obligatory, it serves as a commendable foundation for organizing prompts, facilitating Midjourney's comprehension of your creative vision. Proficiency in vocabulary and a comprehensive grasp of terminology significantly contribute to the effectiveness of prompt construction for optimal results with Midjourney.

Conclusion. What’s Next?

Congratulations! You have acquired the skills to set up Midjourney and produce remarkable artwork. This article aims to provide valuable insights as you embark on your journey of learning and utilizing Midjourney AI. While the basics for beginners have been covered, it's essential to note that there are advanced features, commands, and insights into prompt creation that haven't been addressed here. These will be shared in forthcoming posts and books.

Despite my relatively short time with Midjourney—spanning six months—I have amassed a wealth of knowledge, thanks to the Midjourney Discord Community, various Facebook pages, and prompt-building books. Each day unveils new discoveries as I explore diverse techniques to tailor Midjourney to my preferences. I trust that your experience with Midjourney will bring you the same sense of joy and fulfillment that it has afforded me.

Previous
Previous

Next-Level Midjourney