How to Make Consistent Art in Midjourney V 5.2

Future Tech Pilot
14 Nov 202307:04

TLDRThe video outlines a process for creating consistent art in Midjourney V 5.2 by identifying a desired style, using reference images from a movie, and leveraging Midjourney's features like slash tune for style tuning and remix mode for variations. The creator emphasizes the importance of selecting a style, adjusting stylize values, and including keywords like 'anime aesthetic' to achieve a desired look, resulting in a consistent artistic style across different subjects.

Takeaways

  • 🎨 Identify a desired art style by reflecting on personal preferences, such as a fondness for the art style in Digimon the movie.
  • 📸 Gather reference images from the chosen style and upload them to Discord for easy access during the art creation process.
  • 🔧 Utilize Midjourney's 'slash tune' feature to tune the prompt with the uploaded reference images, combining them with descriptive keywords like 'anime aesthetic'.
  • 📊 Perform a tuning test with 16 style directions, sticking to the default settings initially.
  • 👁️ Evaluate the generated styles and select the one that closely matches your vision, then copy the entire prompt including the image links and style code.
  • 🔄 Experiment with different stylized values (S400 and S40 in the example) to refine the look until the desired outcome is achieved.
  • 🔧 Adjust settings to turn on 'remix mode' and 'high variation mode' to ensure the generated art follows the prompt more closely.
  • 🗑️ Remove the image prompts and style code from the remix box to prevent them from influencing the generation of new subjects.
  • 🏷️ Include specific subjects and descriptive words in the prompt for more accurate and varied results, such as 'full body wide-angle lens action pose'.
  • 🔄 Continuously remix the image that closely matches your desired style to create a series of art pieces with a consistent look.
  • 🎭 Use the process to generate art with a variety of subjects and scenes, maintaining a consistent style throughout.

Q & A

  • What is the main goal of the process described in the transcript?

    -The main goal of the process is to create consistent art in Midjourney V 5.2 by tuning the style using both image prompts and keywords.

  • What personal preference does the speaker have for anime art style?

    -The speaker prefers the art style from the movie Digimon and wishes every anime would look like that.

  • How does the speaker use reference images in Midjourney?

    -The speaker uploads reference images from Digimon the movie to Discord and uses them as a basis for tuning the art style in Midjourney.

  • What is the significance of the 'slash tune' feature in Midjourney?

    -The 'slash tune' feature allows users to fine-tune the art style by copying and pasting image addresses or dragging and dropping images into the prompt.

  • Why is it important to include a keyword when tuning the style in Midjourney?

    -Including a keyword helps to guide the generation of art towards a desired aesthetic, such as 'anime aesthetic' in the speaker's case.

  • How many style directions does the speaker recommend testing?

    -The speaker recommends testing with 16 style directions, as they believe it's sufficient for finding the desired style.

  • What are the two settings that need to be checked in the Midjourney settings to ensure consistent style?

    -The two settings are 'remix mode' and 'high variation mode'.

  • Why is it necessary to delete the image prompts and style code when remixing an image?

    -Deleting the image prompts and style code prevents them from influencing the remix, allowing Midjourney to focus more closely on the user's new subject and desired style.

  • What is the purpose of using a lower stylize value after tuning the style?

    -Using a lower stylize value helps Midjourney follow the user's prompt more accurately, resulting in a more consistent style.

  • How can the process described in the transcript be repeated for different characters or scenes?

    -After finding a style that closely resembles the desired look, the user can continue remixing from that point, changing the subject or scene as needed to maintain a consistent style.

Outlines

00:00

🎨 Achieving Consistent Style in Art with Mid Journey

The paragraph discusses the process of creating art with a consistent style using Mid Journey, a tool that allows users to upload reference images and fine-tune their prompts. The creator shares their personal preference for the art style from the movie 'Digimon' and explains how they used screenshots from the movie as reference images. They guide the user through the steps of uploading images to Discord, using the 'slash tune' feature to tune the prompt, and adjusting the style with both image links and keywords like 'anime aesthetic'. The creator emphasizes the importance of testing different style directions and selecting the one that closely matches the desired look. They also mention the need to adjust settings for 'remix mode' and 'high variation mode' to ensure the final output aligns with the desired style. The paragraph concludes with a demonstration of how to remix the image with a new subject, such as Wonder Woman, to maintain the consistent style.

05:01

🖌️ Refining Art Generation with Remix and Variation

This paragraph delves into the refinement process of art generation using Mid Journey's features. The creator advises on how to exclude certain words from the generation to avoid unwanted vibes, using 'lens' as an example. They express satisfaction with the results, particularly with image number three, and note that the inclusion of 'anime aesthetic' in the prompt can yield varying outcomes. The creator underscores the significance of the stylize value (S40) in achieving the desired look and shares examples of successful remixes with different characters and scenes. They also provide a recap of the workflow, which involves tuning the style with image prompts, selecting the preferred style, generating variations, and deleting unnecessary elements to focus on the subject. The paragraph ends with an encouragement to continue remixing to achieve a consistent style across generations and a suggestion to consult the video for further guidance on tuning and workflow.

Mindmap

Keywords

💡Midjourney V 5.2

Midjourney V 5.2 is a reference to a specific version of a digital art creation tool that the speaker is using to generate images. It is the platform where the process of creating consistent art style is being discussed and applied. The video revolves around the features and functionalities of this version to achieve a desired aesthetic outcome.

💡Style Identification

Style Identification is the process of recognizing and selecting a particular artistic style that one wishes to emulate or achieve. In the context of the video, the speaker identifies a specific anime aesthetic from their childhood, inspired by Digimon the movie, as the desired style for their art.

💡Reference Images

Reference images are visual examples that serve as a guide or inspiration for the art being created. They are used to help the artist or the AI understand the visual elements and aesthetic they are aiming for. In the video, the speaker gathers screenshots from Digimon the movie to use as reference images for the desired art style.

💡Slash Tune

Slash Tune is a feature in Midjourney V 5.2 that allows users to fine-tune their prompts to achieve a specific look or style. It enables the user to adjust the AI's output to better match their desired aesthetic by using image links and keywords.

💡Style Test

A style test is a set of generated images created by the AI to showcase different stylistic variations based on the input provided by the user. It helps the user to select the most appropriate style direction for their project.

💡Remix Mode

Remix Mode is a setting within Midjourney V 5.2 that, when turned on, allows the AI to create new images based on a previously generated image, rather than strictly following the original prompt. This feature helps in achieving consistency in style and aesthetic across multiple images.

💡Variation Button

The variation button is a tool that, when clicked, opens a remix box which allows users to modify the elements of a previously generated image. This feature is used to further fine-tune the output and achieve the desired look.

💡Stylize Value

The stylize value is a parameter in Midjourney V 5.2 that controls the degree to which the AI applies the chosen style to the generated image. Adjusting the stylize value can help achieve the desired balance between the original content and the applied style.

💡Consistent Style

Consistent style refers to the ability to generate a series of images that have a uniform look and feel, which is crucial for creating a cohesive visual narrative or collection. The video provides a step-by-step process to achieve this consistency using Midjourney V 5.2.

💡Image Prompts

Image prompts are visual inputs that guide the AI in generating images with specific characteristics or elements. They are used in conjunction with text prompts to provide a more detailed direction for the AI's output.

💡Aesthetic

Aesthetic refers to the visual style or appearance of the art, which includes elements such as color, composition, and mood. In the video, the speaker is aiming for an anime aesthetic inspired by the art style of Digimon the movie.

Highlights

Creating consistent art style in Midjourney V 5.2 involves a process that starts with identifying a desired style.

The user's personal preference for art style was influenced by the movie 'Digimon' from their childhood.

Midjourney's new feature 'slash tune' is utilized to achieve the desired art style by tuning the prompt with reference images.

Reference images from 'Digimon' movie were uploaded in Discord to access easily during the tuning process.

A combination of image links and keywords like 'anime aesthetic' is recommended for the tuning process.

A tuning test is created with 16 style directions, but it's suggested not to exceed this number.

The user demonstrated how to run the same prompt with different stylized values to refine the desired look.

Settings should be adjusted to turn on 'remix mode' and 'high variation mode' for better results.

The image prompts and style code should be removed in the remix box to avoid affecting the new subject.

A lower stylize value is suggested for closer adherence to the prompt, ensuring consistency in style.

The process allows for the generation of art with a consistent style, as demonstrated with various characters like Wonder Woman and the Hulk.

The importance of the stylize value (S40) was emphasized for achieving the desired outcome.

The user shared their satisfaction with the results, which evoked the emotion and feel they were aiming for.

The method can be applied to everyday scenes, not just characters, to maintain the consistent style.

The user stressed the effectiveness of the tuning and remixing workflow for achieving a consistent art style.

The video provides a comprehensive guide for users who may be confused about the tuning process and workflow.

The user demonstrated how to achieve a consistent style across different subjects and scenes, showcasing the versatility of the method.