23 NEW Midjourney Tips because I like you

Future Tech Pilot
7 Oct 202423:28

TLDRThis video offers 23 creative tips for enhancing your Midjourney experience. From reimagining objects as something else to pixel art techniques, the tips cover a range of imaginative prompts. Learn how to adjust quality settings for better images, personalize your creations with unique codes, and explore different aspect ratios for unique compositions. The video also introduces a 'Shape of X' trick, offers advice on managing your create page, and inspires with a 'Style Safari' website for discovering new styles. Dive into these tips to push your creative boundaries with Midjourney.

Takeaways

  • 🔄 Use the word 'reimagined' in your prompt to transform one thing into another, such as an animal into another animal.
  • 🎨 For pixel art consistency, use style reference with permutations and adjust style weight parameters.
  • 🖥️ Check your zoom level on the website to ensure all options are visible, like aspect ratio changer.
  • 🌐 Images created on Discord show up on the website, but not vice versa. Use job ID to transfer them.
  • 📹 Experiment with video grids by generating on Discord and using as input for Luma.
  • 📈 Request higher quality images by adding 'D-q' followed by a number to your prompt, though it costs more.
  • ⏳ Be aware that fast hours now expire after 60 days and require an active subscription.
  • 🃏 Create unique playing cards by editing the center of a card design while keeping the border.
  • 📐 Use 'arranged' in your prompt to create compositions with geometric shapes like triangles or spirals.
  • 🧑‍💻 Keep track of your personalization codes with the '/list p' command on Discord.
  • ✂️ Edit images one element at a time for better results in the editor.
  • 🎁 Create a GIF by expanding a square canvas into a wider aspect ratio and cycling through different effects.
  • 📏 Try different aspect ratios like 3x1 or 1x3 for unique compositions.
  • 🔍 Add 'd-p' to a prompt to run it with your personalization code for unique results.
  • 🧹 Hide individual images on the create page by going into the image options and selecting hide.
  • 🌟 Experiment with style reference codes to achieve unique textures and styles.
  • 🔢 Prompt for a 3D shape combined with a subject in a square aspect ratio for interesting results.
  • ⚖️ Adjust the strength of personalization code and style reference code with additional parameters.
  • 🎮 Try an Unreal Engine render with a low stylize value for early 2000's computer graphics nostalgia.
  • 🔎 Discover new styles by saving random style reference images and asking AI for possible names.
  • 🧝‍♂️ See your character in a specific style by using a style reference and adjusting character weight parameters.
  • 🌐 Explore a vast catalog of style reference codes on the Style Safari website.
  • 🖼️ Experiment with depicting a person at different ages using character reference and adjusting parameters.

Q & A

  • What is the first tip mentioned in the video for using Midjourney effectively?

    -The first tip is to use the word 'reimagined' in your prompt to transform one thing into another, such as a couch into a fridge or an animal into another animal.

  • How does the pixel art workflow suggested in the video help in generating consistent pixel art?

    -The workflow helps by dragging a generated 8-bit pixel art into the prompt box, selecting 'style reference', and using permutations to run all prompts at once. Additionally, adjusting the style weight parameter (D-ssw) to 1000 maximizes the influence of the reference image on the output.

  • What is the purpose of using the aspect ratio changer in Midjourney as mentioned in the video?

    -The aspect ratio changer is used to manually adjust the output image's dimensions. The video suggests that if certain options are missing, adjusting the zoom level of the webpage might reveal them.

  • How can images created on Discord be viewed on the Midjourney website according to the video?

    -Images created on Discord show up on the website automatically. However, to view images created on the website on Discord, one can use the '/for show' command followed by the job ID, which can be copied from the website.

  • What is the video grid trick discussed in the video and how does it work?

    -The video grid trick involves generating a grid of images on Discord or using the website generations, saving the entire grid as one picture, and using that as input for Luma, a video generator. This can create interesting video effects.

  • How can one control the quality of images generated by Midjourney as per the video?

    -One can control the quality by adding 'D-q' followed by a number to the end of the prompts. A higher number, like 2, increases the generation time for higher quality, while a lower number, like 5, decreases quality for faster and cheaper generations.

  • What change has Midjourney made regarding 'fast hours' according to the video?

    -Starting October 1st, 'fast hours' purchased will expire after 60 days, unlike before where they stayed with the user indefinitely.

  • How can one create personalized playing cards using Midjourney as suggested in the video?

    -One can create personalized playing cards by finding a card design, using the editor to select and change the middle part of the card while keeping the border, and then using the 'very strong' feature to find more results fitting the new style.

  • What is the 'Arrangement' trick mentioned in the video for adding structure to compositions?

    -The 'Arrangement' trick involves picking a subject and a scene, then arranging them in a chosen shape like triangles, circles, or spirals to add geometry and structure to the composition.

  • How can one keep track of their personalization codes in Midjourney as discussed in the video?

    -One can use the '/for list p' command on Discord to bring up every code they've ever created, helping to keep track of their personalization codes.

Outlines

00:00

🎨 Creative Prompting Techniques

The paragraph introduces various tips for enhancing AI-generated art. It starts with the creative use of the word 'reimagined' in prompts to transform one object into another, such as a couch into a fridge. The speaker credits the super user Club on Discord for this trick. Examples include a walrus reimagined as a lion and a husky as a bird. The paragraph also discusses generating pixel art with consistency by using style references and adjusting style weight parameters. The speaker, Nolan, shares additional tips on using aspect ratio, syncing Discord with the website, and using video grids for a unique effect. The segment ends with a call to action for likes and mentions of Patreon support.

05:01

🖼️ Enhancing Image Quality and Managing Fast Hours

This paragraph discusses methods to improve the quality of AI-generated images by using the 'D-q' parameter, which allows for higher quality at the cost of more fast hours. The speaker compares different quality levels and their upscaled versions, suggesting that upscaling doesn't significantly change the quality but the quality level itself does. The paragraph also addresses a change in mid Journey's policy where fast hours now expire after 60 days, prompting users to manage their resources accordingly. Additionally, a trick for creating personalized playing cards using the editor is shared, along with a method to create unique card sets by changing elements within a selected border.

10:03

🌐 Geometric Compositions and Personalization

The speaker introduces the concept of 'arranged' in prompts to create geometric compositions, such as arranging subjects in triangles or circles. This method adds structure to compositions and can yield visually striking results. The paragraph also discusses the importance of keeping track of personalization codes, especially when ranking images, and how to list all created codes on Discord. The speaker shares a workflow for editing images one element at a time for better results and suggests creating gifts by expanding a square composition into a wider canvas, using a higher chaos value for interesting effects.

15:04

🌠 Exploring Aspect Ratios and Discovering Unique Styles

The paragraph focuses on the aesthetic appeal of stretched aspect ratios like 3x1 and 1x3, which can create dramatic compositions. The speaker shares a workflow for personalizing styles by adding 'd-p' to prompts, ensuring results are unique to the user's style. There's also a tip on cleaning up the create page by hiding individual images. The speaker highly recommends a specific style reference code for its ability to portray texture and layers, and encourages experimentation with different aspect ratios and subjects. The paragraph concludes with a method to create a GIF by uploading varied images to a site like easy GIF.

20:04

🔍 Adjusting Style Influence and Nostalgic Rendering

This paragraph discusses techniques to adjust the influence of personal style and style reference codes in prompts. The speaker explains how to use parameters like 'D-s' and '--ssw' to control the strength of personalized style and style reference, respectively. A live example is given using a 'Scorch' code to demonstrate the significant visual differences achieved by altering these parameters. The paragraph also explores creating nostalgic early 2000's computer graphics by using 'Unreal Engine render' with a low stylize value and 'nii Journey' as the version, evoking memories of early computer graphics.

🔑 Unlocking New Styles and Character Transformations

The speaker shares a unique workflow for discovering new styles by using 'srf random' and saving the results to inquire about style names from GPT, leading to creative and inspiring style names. The paragraph also covers a method for transforming a character into another in a specific style by using the original character as a style reference and adjusting parameters to prioritize the style. The speaker suggests a website called 'Style Safari' as a comprehensive catalog of style reference codes. The paragraph concludes with a technique called 'lineage' for generating a character across different ages, using terms like 'newspaper photo' and adjusting parameters to achieve an aged look.

Mindmap

Keywords

💡Reimagined

The term 'reimagined' refers to the process of creating a new version or interpretation of something that already exists. In the context of the video, it is used to describe how the AI tool can transform one object or animal into another, such as turning a couch into a fridge or a walrus into a lion. This showcases the AI's ability to handle abstract concepts and create unique imagery based on the user's prompts.

💡Pixel Art

Pixel art is a form of digital art where images are created using small, pixel-sized blocks of color. The video discusses a workflow for generating consistent pixel art using the AI tool, emphasizing the use of style reference and style weight parameters to achieve the desired 8-bit aesthetic. This technique allows users to create retro-style images that mimic the graphics of early video games.

💡Aspect Ratio

The aspect ratio is the proportional relationship between the width and the height of an image or screen. The video mentions that the aspect ratio changer might not be visible if the user is zoomed in too much on the website, suggesting that users adjust their zoom level to access this feature. The aspect ratio is crucial for构图 and ensures that the generated images fit the desired dimensions.

💡Discord

Discord is a communication platform mentioned in the video where users can create and share images. The script explains how to transfer images from the AI tool's website to Discord using a job ID, allowing for further interaction and sharing within Discord communities. This highlights the collaborative aspect of using AI art tools within social platforms.

💡Video Grids

Video grids refer to the arrangement of multiple images or frames in a grid pattern, often used in video editing or animation. The video describes a trick for using video generators with AI-generated grids, where the entire grid is saved as one picture and used as input for another AI tool. This process can create dynamic and varied视觉效果, as demonstrated by the examples provided.

💡Quality

In the context of the video, 'quality' refers to the resolution and detail of the AI-generated images. The script explains that users can request higher quality images by adding a 'D-q' parameter followed by a number to their prompts, with higher numbers indicating better quality but also increased generation time and cost. This allows users to balance between image quality and the resources they are willing to expend.

💡Personalization Code

The personalization code is a unique identifier used in the AI tool to store and apply a user's preferences and style choices across different image generations. The video mentions that this code changes每次用户对图像进行排名,which can make it difficult to keep track of preferences. Using a command like 'for/list P' on Discord can display all the codes a user has created, helping to manage and reuse personalized styles.

💡Editor Workflow

The editor workflow refers to the process of refining and fixing elements within an image using the AI tool's editing features. The video advises fixing issues one at a time rather than attempting to correct multiple areas simultaneously, as this increases the likelihood of achieving the desired outcome. This approach helps users to make incremental improvements to their images, ensuring greater control over the final result.

💡Gift

In the script, 'gift' likely refers to creating animated images or GIFs using the AI tool. The video suggests creating a base image with a square aspect ratio and then expanding it to a wider canvas to create a 'gift' by altering the surrounding elements. This technique allows for the creation of dynamic and interesting visual effects that can be循环 through different scenes or expressions.

💡Aspect Ratios

Aspect ratios are discussed again in the context of creating stretched or compressed compositions. The video mentions using tall or wide aspect ratios like 3x1 or 1x3 to create dramatic and visually interesting images. This approach forces the AI to adapt the subject matter to the canvas shape, resulting in unique and artistic interpretations.

💡Style Reference Code

A style reference code is a specific set of parameters that defines a particular visual style for the AI to follow when generating images. The video highlights a particular style reference code (36101, 37247) that produces grungy textures with a mix of white, black, and orange colors, demonstrating how certain codes can create distinctive and appealing visual effects.

Highlights

Reimagined word in prompts can transform one object into another, such as a couch into a fridge.

Animal to animal transformations work best with the reimagined concept.

Pixel art can be generated consistently by using style reference and adjusting style weight parameters.

Aspect ratio options may be hidden due to zoom levels on the website interface.

Images created on Discord do not automatically show up on the website, but can be accessed via job ID.

Video grids can be generated by using Discord first and then inputting into Luma Dream Machine.

Higher quality images can be generated by adding D-q and a number to the end of prompts.

Fast hours now expire after 60 days, requiring an active subscription to use.

Playing cards can be personalized by editing the card's center while keeping the border.

Arrangement in prompts can create unique compositions with geometric structures.

Personalization codes change with each ranking, and can be listed using the for/list P command on Discord.

Editing images should be done one issue at a time for the best results.

Creating GIFs from a series of images can be done using a square aspect ratio and uploading to easy GIF.

Stretched aspect ratios like 3x1 and 1x3 can produce unique and interesting compositions.

Using d-p with personalization code ensures results are uniquely styled.

Individual images on the create page can be hidden through the image options.

Style reference code 36101 37247 is particularly effective for portraying grungy textures.

3D shapes combined with subjects can create visually striking art pieces.

Adjusting the strength of personalization and style reference codes can significantly alter image outcomes.

Unreal Engine render with a low stylize value can recreate early 2000's computer graphics.

Discovering new styles can be aided by using GPT to generate names for unidentified styles.

Characters can be portrayed as other characters in a specific style using style reference and character reference.

Style Safari is a comprehensive website for exploring different style reference codes.

Depicting a person in different ages can be achieved by adjusting character weight and stylize value.