Best Character Consistency Midjourney Workflow!

Snowball AI
18 Mar 202407:41

TLDRIn this video, the presenter shares their excitement about a new feature in Snowball AI called 'character reference', which is designed to enhance character consistency across multiple images generated by AI. The feature works similarly to the 'style reference' feature, allowing users to input an image of their character and maintain consistency in appearance, including facial features, hair, and clothing. The presenter explains how to use the feature with the 'D-C' command and adjust the character weight to focus on specific aspects of the character. They also highlight that the feature works best with characters created in Midjourney and can be combined with the style reference feature for creating a series of images with a consistent character and style. The video demonstrates the process of generating a character design sheet, using the character sheet as a reference for new images, and combining character and style references to achieve a consistent look. The presenter also discusses common issues, such as handling multiple characters in an image and adjusting the style to match the character. They conclude by showcasing the final illustrations created using Midjourney and Photoshop, emphasizing the importance of lighting and the potential of these tools for creating detailed and consistent character illustrations.

Takeaways

  • 🎨 The video discusses the use of a new 'character reference feature' in Midjourney, which helps to create consistent character illustrations across multiple images.
  • 📈 The 'character reference feature' is similar to the 'style reference feature' and allows for maintaining consistency in a character's face, hair, and clothing.
  • ✅ By using the character reference feature with a specific syntax (-DC followed by an image URL), one can control the reference strength with a character weight (CW) value between 100 and 0.
  • 🚫 The feature is not designed for real people or photos and works best with characters created within Midjourney.
  • 🧩 It's suggested to use multiple images as references, especially when depicting different angles or poses of the character.
  • 🖼️ The video demonstrates creating a character design sheet by generating various poses and expressions of a character, then using these as references for further illustrations.
  • ✂️ It's recommended to crop and edit each image to fix deformities before using them as references for better results.
  • 🖌️ The video also shows how to combine the character reference with the style reference feature to maintain both character and stylistic consistency.
  • 💡 The presenter shares a trick of using Photoshop for editing the background and making minor adjustments to the character for a more polished illustration.
  • 🔍 The 'zoom out' feature in Midjourney is highlighted as a way to adjust the framing of the character from close-up to full body shots.
  • 🌟 Lighting is emphasized as an important aspect of the final illustration, which can be adjusted in Photoshop for a more realistic effect.
  • 🔗 The video concludes with a recommendation for a Photoshop tutorial link for beginners, focusing on use cases similar to those shown in the video.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is about creating consistent characters using the latest Midjourney feature, specifically the new character reference feature.

  • What has been a challenge when using generative AI for character consistency?

    -A huge issue with generative AI has been the inconsistency in character appearances due to the AI's nature to give random results even with detailed character descriptions.

  • How does the new character reference feature in Midjourney work?

    -The character reference feature works by using a specific syntax ('--v 5 --ar 16:9 --stylize 500 --q 2 --v 5 --ar 16:9 --stylize 500 --q 2 --test'). It allows the replication of the face, hair, and even clothes of a character across multiple images, with the ability to adjust the reference strength using the '--CW' parameter.

  • What is the default character weight in the character reference feature?

    -The default character weight is 100.

  • What happens if you set the character weight to zero?

    -If you set the character weight to zero, the AI will just focus on the face, which is useful for changing outfits, hair, and adding accessories to characters.

  • What is the limitation of the character reference feature when it comes to details like logos on a t-shirt?

    -The character reference feature may not be able to copy every detail, such as a specific logo on a t-shirt, in the new image it creates.

  • How can the character reference feature be used to draw inspiration from multiple images?

    -The character reference feature allows users to upload multiple images of a character to create an image where the character is in a different pose or angle, like facing away from the viewer.

  • What is the recommended aspect ratio for creating a character sheet in Midjourney?

    -The recommended aspect ratio for creating a character sheet is 16:9, which provides a horizontal image with room for more outputs of the character.

  • Why is it suggested to crop each image separately and fix deformities before saving when using character images as reference?

    -Cropping each image separately and fixing deformities increases the chances of getting better results when using these images as reference, especially for main characters in stories with more than 10 illustrations.

  • How can the style reference feature be combined with the character reference feature?

    -The style reference feature can be combined with the character reference feature by adding the '--SRE' parameter followed by the URL or link to the character sheet image, ensuring that the style is copied along with the character into the new image generated.

  • What is the role of Photoshop in enhancing the consistency and quality of the generated images?

    -Photoshop is used to edit the background, add minor lighting effects, fix details such as removing backgrounds, adding missing elements like hands, and adjusting the overall appearance to match the desired style and consistency.

  • What is the zoom out feature in Midjourney and how is it used?

    -The zoom out feature in Midjourney allows users to expand the view of the generated image to show more of the character's body. It is used when the initial images are close-ups or medium close shots, and a full-body view is desired.

Outlines

00:00

🎨 Character Consistency with M Journey's New Feature

The speaker is excited to explore M Journey's latest feature, the character reference, which aims to address the challenge of creating consistent characters across multiple images. They discuss the previous issue of AI models generating random results despite detailed character descriptions. The video demonstrates using the new feature by creating a character design sheet with various poses and expressions, and then using these images as references to generate new images with consistent character appearances. The process involves writing prompts, using the character reference feature with a specific syntax (D-C followed by an image URL), and adjusting the character weight (CW) to focus on different aspects of the character. The speaker also covers tips for refining the process, such as cropping images and using multiple references for different angles, and combining the character reference with the style reference feature for a cohesive look in a series of images.

05:02

🖼️ Enhancing Character Images with Photoshop and M Journey

The speaker continues by showing how to refine the generated images using Photoshop for tasks like background editing, character extraction, and lighting adjustments. They also discuss the challenge of handling multiple characters in a single image and how Photoshop can be used to address this. The video demonstrates creating a new image of the character watching a movie and eating popcorn, using both character and style references to maintain consistency. The process includes using AI to select and edit the character, generating a new background, and adjusting lighting to create a realistic scene. The speaker also shares a trick for using M Journey's zoom out feature to get full-body shots and how to use Photoshop to fix areas like missing limbs. They conclude by recommending a Photoshop tutorial for beginners and encourage viewers to subscribe.

Mindmap

Keywords

💡Character Consistency

Character consistency refers to the ability to maintain a character's visual appearance and personality traits across different instances within a story or series of images. In the context of the video, it is a crucial aspect when using generative AI to create illustrations, as it ensures that the same character is depicted with similar appearances in multiple images, which is essential for creating a cohesive narrative or series.

💡Midjourney

Midjourney, in the context of the video, refers to an AI model that is used for generating images. It is mentioned as the best AI model for this purpose by the speaker, highlighting its significance in creating consistent characters across different images. The term 'Midjourney' here is likely a proper noun referring to a specific AI tool or platform.

💡Character Reference Feature

The character reference feature is a new tool within the Midjourney AI that allows users to input an image of a character and generate new images that maintain the same character's appearance. It is designed to address the challenge of character consistency in generative AI. The feature is used by typing 'D-C' followed by an image URL, and it can replicate facial features, hair, and clothing, making it a powerful tool for creating character design sheets and narratives.

💡Style Reference

Style reference is a feature similar to character reference that allows users to maintain a consistent artistic style across various images. It is used by providing an image as a reference point, and the AI generates new images that adhere to the style of the provided reference image. In the video, it is combined with the character reference feature to create images that are both stylistically and character-consistent.

💡Character Weight (CW)

Character weight (CW) is a parameter that can be adjusted when using the character reference feature. It allows users to modify the strength of the character reference, with a range from 100 (strongest reference) to 0 (weakest, focusing only on the face). This feature is useful for making subtle changes to a character's appearance, such as changing outfits or hairstyles, without altering the character's identity.

💡Children's Book Illustrations

Children's book illustrations are visual representations used in children's books to depict characters, scenes, and narratives. In the video, the speaker uses the Midjourney AI to create character sheets in the style of children's book illustrations, which are typically characterized by their cartoonish and 2D appearance. These illustrations are important for creating a visual narrative that complements the text in children's books.

💡Photoshop

Photoshop is a widely used software for image editing and manipulation. In the video, it is mentioned as a tool for refining the images generated by the Midjourney AI. The speaker uses Photoshop to crop images, fix deformities, adjust lighting, and combine elements to create final illustrations. It is an essential tool for post-processing AI-generated images to achieve a polished and professional look.

💡Aspect Ratio

The aspect ratio is the proportional relationship between the width and the height of an image or screen. In the video, the speaker mentions using an aspect ratio of 16:9, which is a common widescreen format that allows for more horizontal space and is suitable for creating character design sheets with multiple outputs.

💡Generative AI

Generative AI refers to artificial intelligence systems that can create new content, such as images, music, or text, based on existing data or instructions. In the context of the video, generative AI is used to generate consistent character images and is a key technology for creating illustrations and visual content without manual drawing.

💡URL

URL stands for Uniform Resource Locator, which is a reference to a web resource that specifies its location on a computer network and a mechanism for retrieving it. In the video, the speaker instructs how to use the URL of a character sheet image as a reference for the Midjourney AI to generate new images with the same character, highlighting the importance of the URL in the process of image generation.

💡Lighting

Lighting in the context of image editing refers to the manipulation of light sources and shadows to enhance the mood, depth, and realism of an image. The video emphasizes the importance of lighting in creating visually appealing illustrations. The speaker demonstrates how to adjust lighting to make the character appear as if they are in a cinema or a spaceship, adding a layer of realism and immersion to the final image.

Highlights

Introduction of a new character consistency feature in Midjourney AI for creating consistent characters across multiple images.

The character reference feature is similar to the style reference feature, allowing for consistent character appearances.

Character weight (CW) can be adjusted from 100 to 0 to control the level of detail in character replication.

The feature is not designed for real people or photos and works best with characters created in Midjourney.

Combining character and style reference features can create a series of images with consistent style and character.

Using multiple images as reference can help create images of characters in different poses or angles.

Cropping and editing images separately can improve the quality of character consistency.

Specific images of a character can be used for specific scenes to maintain consistency.

An example of generating a character in multiple poses and angles to create a character design sheet.

Combining the new character reference with Photoshop for creating beautiful illustrations.

The importance of adjusting the aspect ratio for generating character images.

Using the zoom out feature in Midjourney to show more of the character's body in the images.

Editing the background and adding minor lighting effects to enhance the character images.

The process of generating a new background and combining it with the character using style references.

Tips on using lighting and blending modes in Photoshop to improve the final illustration.

The use of generative fields in Photoshop for fixing specific areas of the character image.

Recommendation of a Photoshop tutorial for beginners focused on use cases similar to those shown in the video.

Emphasis on the capability of these tools to create exactly what the user wants when used properly.