midjourney v6 advanced tutorial ,MJ同一场景出现多个一致性角色的绘图方法,局部修改,改照片换脸,AI连续角色创作方法,AI绘图如何写提示语

17 Apr 202426:50

TLDRThis tutorial explores advanced techniques in AI art creation, focusing on generating multiple consistent characters in a scene using the CREF parameter in midjourney v6. It covers local modifications, photo face swaps, and maintaining stylistic consistency across images. The host shares tips on crafting effective prompts for AI drawing, demonstrating the process of creating and refining images of an Asian beauty and an American man in various settings, and concludes with anime character examples, highlighting the versatility and power of AI in customizing and unifying character appearances.


  • 😀 The video tutorial introduces a method to draw multiple consistent characters in the same scene using a new function called CREF.
  • 🎨 It demonstrates how to modify parts of an image, such as changing faces in photos, using AI tools for drawing.
  • 🤖 The AI's ability to generate consistent characters across different scenes is enhanced by using specific prompt words and parameters.
  • 📚 The tutorial suggests subscribing to Xiaoyuer's technical school for more drawing lessons and tips.
  • 🔗 The importance of saving images and their links is emphasized to avoid loss due to potential issues with shared links or services.
  • 🖼️ The process of selecting satisfactory images and enlarging them to U1 for better detail is covered.
  • 📝 The script mentions using a knowledge planet platform where all prompt words used by Xiaoyuer are shared.
  • 🌐 A shared rental link for AI services is provided in the description for those interested in trying the tools mentioned.
  • 🎭 The tutorial includes steps for changing character appearances in different scenes while maintaining consistency, using tools like GPT for prompt generation.
  • 🔧 The use of the CREF parameter is highlighted for face-changing in AI drawings, with a link provided for a detailed tutorial on its use.
  • 🌐 The video mentions the potential need to wait for image processing in shared AI services due to high demand.

Q & A

  • What is the main topic discussed in the tutorial?

    -The main topic discussed in the tutorial is the method of drawing multiple consistent characters in the same scene using the midjourney v6 advanced tutorial, including local modifications, changing photos by swapping faces, and AI continuous character creation methods.

  • What is the CREF parameter mentioned in the script, and how is it used?

    -The CREF parameter is a new function in midjourney that allows the appearance of multiple characters in a scene to be consistent. It is used by adding a picture link and the CREF parameter in the prompt to ensure the characters look the same across different scenes.

  • How does the tutorial suggest maintaining consistency in photo styles?

    -The tutorial suggests maintaining consistency in photo styles by using the same camera film, specifically Kodak Portrait 400, in the prompt to ensure that the style of the generated images is unified.

  • What is the significance of saving images and their links as mentioned in the script?

    -Saving images and their links is important to prevent loss of work, especially when using shared links, which may disappear at any time. It also allows for easy access and modification in future projects.

  • How can the GPT tool assist in creating prompts for different scenes?

    -The GPT tool can assist by generating multiple prompts based on specific requirements, such as different scenes and activities for the characters, ensuring that the characters' appearance remains consistent across these scenes.

  • What is the purpose of using the lasso tool in the character modification process?

    -The lasso tool is used to select specific areas of an image for modification, such as the face or hair, allowing for precise changes to be made to the character's appearance without affecting the rest of the image.

  • How does the tutorial address the issue of shared account limitations?

    -The tutorial acknowledges that shared account users may experience wait times due to system busyness and suggests being patient and submitting requests multiple times if necessary.

  • What is the role of the 'Remix' function in the character modification process?

    -The 'Remix' function is used to reassemble parts of an image after modifications. It is activated through the settings and allows for the integration of changes made to individual characters into a single, cohesive scene.

  • How can the CW parameter be used in conjunction with the CREF parameter?

    -The CW parameter, which stands for 'Control Weight', is used in conjunction with the CREF parameter to control the intensity of the modifications. A higher CW value can make the changes more pronounced.

  • What advice does the tutorial give for dealing with delivery failures during the image modification process?

    -The tutorial advises to simply click and submit multiple times in case of delivery failures, as this could be due to system busyness. Persistence in resubmitting the request is recommended.



🎨 Introduction to Character Illustration Technique

The speaker, Xiao Yu'er, introduces a new function to create illustrations of characters that appear identical in different scenes. Despite the need for further optimization, the accomplishment is celebrated. The process involves using a CREF parameter and sharing the experience with the audience. The tutorial covers selecting characters, using specific prompt words, and ensuring consistency in style by simulating a Kodak Portrait 400 film shot. The speaker also promotes their technical school and encourages viewers to subscribe and engage with the content.


🖌️ Step-by-Step Guide to Creating Consistent Character Scenes

Xiao Yu'er provides a detailed guide on creating scenes with consistent characters. The process includes generating character images, using a GPT tool to write prompts, and ensuring a unified style across different scenes. The speaker demonstrates how to use specific prompts to maintain character appearance and suggests using camera film references for style consistency. The tutorial also covers troubleshooting, such as dealing with unsatisfactory results and making multiple attempts to achieve the desired outcome.


🔧 Advanced Character Editing with CREF Parameters

The paragraph explains how to use an updated CREF parameter for advanced character editing in illustrations. Xiao Yu'er guides the audience through opening the Remix function, using the lasso tool for selection, and entering new parameters to modify character features. The process is iterative, with attention to detail to correct minor flaws and achieve a high-quality result. The speaker emphasizes patience and careful reading to master the technique.


🌟 Fine-Tuning Character Details for Perfection

Xiao Yu'er focuses on fine-tuning character details, such as hair and facial features, to achieve a perfect illustration. The process involves selecting the appropriate areas with the lasso tool, adjusting the CREF parameter, and ensuring consistency in character appearance across different scenes. The speaker highlights the importance of selecting a wide enough area for detailed changes and provides tips for dealing with common issues like hair symmetry.


🎭 Experimenting with Anime Character Integration

The speaker discusses the process of integrating anime characters into scenes while maintaining a consistent style. Xiao Yu'er uses descriptive words to define the characters and their environment, then applies the CREF parameter to change facial features and hairstyles. The tutorial includes tips for dealing with complex features like hair and eyes, and the importance of saving work to avoid loss due to shared access issues.


📈 Cost-effective Solutions for AI Art Generation

Xiao Yu'er concludes the video script by discussing cost-effective solutions for using AI tools like Miji for art generation. The speaker compares the prices of different plans and payment options, highlighting the benefits of annual payments and shared access. The tutorial encourages viewers to subscribe to the channel for more content and to share their experiences and suggestions in the comments section.



💡CREF parameter

The CREF parameter is a tool within AI image generation software that allows for the consistent representation of characters across different scenes or images. In the video, it is used to ensure that when multiple versions of the same character appear in various scenes, they maintain a uniform appearance. The script mentions using a CREF parameter to achieve this consistency, indicating its importance in creating a seamless narrative in AI-generated art.

💡AI continuous character creation

AI continuous character creation refers to the process of generating a character in AI art that remains consistent across multiple images or scenes. The video discusses techniques for ensuring that characters look the same in different situations, which is crucial for storytelling and maintaining viewer immersion. The script provides insights into how AI parameters can be manipulated to achieve this, such as using the CREF parameter.

💡Kodak Portrait 400

Kodak Portrait 400 is a type of film known for its color reproduction and grain structure, often used in photography for portrait shots. In the context of the video, it is used metaphorically to describe a style or filter applied to AI-generated images to give them a consistent and specific visual aesthetic. The script mentions adding 'Kodak Portrait 400 film shot' to the prompt words to maintain a consistent style across images.

💡Prompt words

In AI image generation, prompt words are the textual descriptions or commands that guide the AI in creating an image. They are crucial for directing the style, content, and mood of the generated artwork. The video script discusses the importance of carefully choosing prompt words to generate specific characters and scenes, such as describing an 'Asian beauty' or an 'American boy'.

💡Face-changing tutorial

A face-changing tutorial in the context of AI art refers to a guide on how to alter the facial features of a character in AI-generated images. The video script mentions a tutorial that explains the use of the CREF parameter for changing faces, indicating a specific technique for modifying character appearances in AI art.

💡Remix function

The Remix function, as mentioned in the script, is a feature within AI image generation tools that allows users to modify existing images or elements within them. It is part of the process of refining AI-generated content to better fit the user's vision, such as changing the face of a character to match a specific style or appearance.


GPT, or Generative Pre-trained Transformer, is a type of AI language model capable of generating human-like text based on given prompts. In the video, GPT is used to help write prompt words for AI image generation, showcasing its utility in creative processes beyond text generation, such as assisting in the conceptualization of scenes and characters.

💡Anime characters

Anime characters are stylized characters from Japanese animated productions. The video script discusses the creation of anime-style characters using AI, emphasizing the ability to generate consistent and specific appearances, such as characters with 'silver hair' or 'pink hair,' and placing them in various scenes like a park.


Miji, in the context of the video, seems to refer to a platform or tool used for AI image generation. The script mentions using Miji for generating images and suggests that it has different pricing plans and features, indicating that it is a service that users can subscribe to for creating AI art.

💡Knowledge Planet

Knowledge Planet, as mentioned in the script, appears to be a platform or community where the video creator shares additional resources, such as prompt words and tutorials. It is a place for those interested in AI image generation to access more in-depth information and learn from the creator's experiences.


Introduction of a new function to achieve character consistency in AI-generated images using the CREF parameter.

Tutorial on creating two characters that look exactly the same in different scenes using AI.

Explanation on how to use the CREF parameter for character consistency in AI image generation.

The importance of using the same camera film prompt to maintain style consistency in images.

Demonstration of generating an Asian beauty and an American man with specific clothing using AI prompts.

Utilization of Kodak Portrait 400 film prompt for style consistency in image generation.

Method of selecting and saving satisfactory AI-generated images for later use.

Technique for changing the background and perspective of AI-generated characters in different scenes.

Use of GPT to generate multiple prompts for characters in various settings.

The process of refining AI-generated images for character consistency and facial features.

How to modify specific parts of an AI-generated image using the lasso and rectangle tools.

Explanation of the CW parameter and its role in adjusting the clarity and detail of AI-generated images.

Addressing common issues like delivery failure and waiting times in shared AI services.

Technique for changing facial features and expressions of AI-generated characters using the CREF parameter.

The successful application of the CREF parameter in changing the ethnicity and features of a character's face.

Demonstration of generating anime characters and modifying their features in a consistent style.

Final thoughts on the power of the CREF parameter in character customization for AI-generated content.

Encouragement for viewers to subscribe and share the tutorial for creating consistent characters in AI art.