How to Make Consistent Art in Midjourney V 5.2
TLDRThe video provides a detailed guide on creating consistent art style in Midjourney V 5.2. The process begins with identifying a desired art style, using personal preferences such as the Digimon movie's art style as an example. The creator shares a step-by-step method involving the use of reference images, the 'slash tune' feature in Discord, and the inclusion of keywords like 'anime aesthetic' to guide the style. The tutorial emphasizes the importance of tuning the style with image links and words, selecting a preferred style, and then remixing the generated images with variations in stylized values. The settings are adjusted to enable 'remix mode' and 'high variation mode' for more diverse outputs. The final step involves deleting image prompts and style codes before inputting a new subject to maintain a consistent style across generations. The video concludes with a recap of the process and an encouragement to experiment with different characters and scenes to achieve the desired art style.
Takeaways
- π¨ **Identify a Desired Style**: Start by knowing the specific art style you want to create, such as the Digimon movie style mentioned.
- π **Gather Reference Images**: Collect screenshots or images that exemplify the style you're aiming for and save them on your computer.
- π **Upload to Discord**: Use the upload feature in Discord to make your reference images easily accessible in a grid format.
- π **Use Midjourney's Slash Features**: Utilize the 'slash tune' feature to customize your prompt with the image links and keywords.
- π **Tune the Style**: Include both image links and descriptive keywords like 'anime aesthetic' in your prompt to guide the style.
- 𧩠**Select Style Directions**: After tuning, choose a style direction that aligns with your vision, possibly using 16 style directions for a test.
- π **Copy and Paste Prompt**: Once a preferred style is selected, copy the entire prompt, including image links, into Discord.
- π **Experiment with Stylized Values**: Run the prompt with different stylized values (e.g., S400, S40) to see variations and find the closest match to your desired look.
- βοΈ **Adjust Settings for Variation**: Ensure 'remix mode' and 'high variation mode' are enabled in the settings for more diverse outputs.
- π« **Remove Image Prompts and Style Code**: When remixing, delete the original image prompts and style code to avoid them influencing the new subject.
- π **Input New Subject**: Introduce a new subject into the prompt, keeping the stylized value and any other relevant keywords.
- π **Continue Remixing**: You can remix the original picture or continue with new subjects to maintain a consistent style across generations.
- π **Iterate and Refine**: Keep iterating the process with different characters or scenes, adjusting the stylized value and keywords as needed to refine the style.
- π **Inclusion of Descriptive Words**: Adding more descriptive words like 'full body', 'wide-angle lens', or 'action pose' can help achieve a more specific look.
- β **Exclusion of Unwanted Keywords**: Exclude words that might introduce an unwanted vibe, ensuring the output aligns with the desired anime style.
Q & A
What is the main goal of the process described in the transcript?
-The main goal is to create a consistent art style in Midjourney V 5.2 by using reference images and the 'slash tune' feature to achieve a desired look, such as an anime aesthetic.
How does one start the process of creating consistent art in Midjourney V 5.2?
-One starts by identifying a specific style they want to achieve, gathering reference images from a source that embodies that style, and uploading those images to Discord.
What is the 'slash tune' feature in Midjourney V 5.2?
-The 'slash tune' feature allows users to fine-tune a prompt by incorporating image links and keywords to guide the AI towards a specific style or aesthetic.
Why is it necessary to include both image links and keywords when using the 'slash tune' feature?
-Including both image links and keywords helps the AI understand the desired style more accurately. Keywords provide a textual description, while image links offer visual references for the AI to emulate.
What is the recommended number of style directions to use when creating a tuning test?
-It is recommended to use 16 style directions for the tuning test, as it provides a good balance without being overwhelming.
How does one select a style they like from the tuning test?
-After the tuning test is created, one can go through the different options, compare the results, and select the style that most closely matches their desired look.
What are stylized values and how are they used in the process?
-Stylized values are numerical settings that control the level of stylization in the generated art. They can be adjusted higher or lower than 100 to fine-tune the style of the artwork.
Why is it important to turn on 'remix mode' and 'high variation mode' in the settings?
-Turning on 'remix mode' and 'high variation mode' allows for more diverse and unique variations of the art, which helps in achieving a consistent style across multiple generations.
What should one do after selecting a preferred style from the tuning test?
-One should copy the entire prompt, including the image links and style code, and run it with different stylized values to see how the style translates with different settings.
How does the 'variation' button help in refining the art style?
-The 'variation' button allows users to make adjustments to the selected style by deleting image prompts and the style code, and then inputting a new subject to generate variations that closely follow the desired aesthetic.
What is the significance of using a lower stylize value in the process?
-Using a lower stylize value ensures that the AI follows the user's prompt more closely, which is crucial for maintaining a consistent style across different generations of art.
Can the process be repeated with different subjects or scenes?
-Yes, the process can be repeated with different subjects or scenes by remixing the original picture or continuing with the new subject, which helps in maintaining a consistent style.
Outlines
π¨ Creating Consistent Art Styles with Mid Journey
The speaker describes a process for creating a consistent art style using the Mid Journey tool. They start by identifying a desired art style, inspired by the movie 'Digimon'. The process involves gathering reference images, uploading them to Discord, and using the 'slash tune' feature to incorporate these images into a prompt. They recommend using keywords like 'anime aesthetic' to refine the style and suggest using 16 style directions for a tuning test. Once a preferred style is chosen, they explain how to adjust the stylized values and use the remix feature to create variations of the desired style without specific characters. The speaker also emphasizes the importance of enabling 'remix mode' and 'high variation mode' for more diverse results.
π Refining and Remixing Art Styles in Mid Journey
The speaker continues to elaborate on the process of refining and remixing art styles in Mid Journey. They explain how to select a style that closely matches the desired look and then use the variation button to create different versions. After choosing an image that fits the preferred style, they advise deleting the image prompts and the style code to avoid unwanted influences on the remix. They also suggest using a lower stylize value to ensure the prompt is closely followed. The speaker then demonstrates how to input a new subject, such as 'Wonder Woman', to create a remix of the chosen image. They highlight the importance of S40 stylized value and the potential for creating consistent styles across different subjects and scenes. The speaker concludes by summarizing the steps for tuning a style with image prompts, making selections, generating styles, and remixing images to achieve a consistent look.
Mindmap
Keywords
Midjourney V 5.2
Consistent Style
Reference Images
Discord
Slash Tune
Anime Aesthetic
Stylize Values
Remix Mode
High Variation Mode
Variation Button
Descriptive Words
Highlights
The process of creating consistent art style in Midjourney V 5.2 involves identifying a desired style and using a combination of new and old features.
The creator was inspired by the art style from the movie 'Digimon' during childhood and aimed to replicate it in Midjourney.
Reference images from the movie were gathered and uploaded to Discord for easy access and as a visual guide.
Midjourney's new 'slash tune' feature is used to customize the style of the generated art.
Image links and keywords are combined in the prompt to guide the style of the art.
A tuning test is recommended with 16 style directions to find the preferred style.
Once a preferred style is selected, the entire prompt, including image links, is copied into Discord.
Different stylized values are tested to fine-tune the look of the generated art.
The prompt in the settings is adjusted with 'remix mode' and 'high variation mode' turned on for more diverse results.
The variation button is used on the preferred image to remix and adjust the art without the original image prompts or style code.
A new subject can be input at the front of the prompt for consistent style across different characters or scenes.
The process can be repeated with different subjects or scenes to maintain a consistent art style.
Descriptive words can be included in the prompt to refine the style and achieve specific looks.
The importance of using a lower stylize value (S40) for consistency is emphasized.
The creator demonstrates the process with examples of Wonder Woman, Hulk, and Iron Man, achieving a consistent anime aesthetic.
The process involves tuning, selection, remixing, and iterating to achieve the desired consistent art style.
The creator expresses satisfaction with the results, noting that the art generated evokes the desired emotion and feel from the Digimon movie.
The final step is to continue remixing the image or start anew with different subjects to maintain a consistent style across generations.