This New AI Tool Does Consistent Characters (For Real)

Theoretically Media
4 Jan 202413:36

TLDRThe video introduces Scenario, an AI image generator that creates consistent characters for AI imagery. The presenter guides viewers through the process of using Scenario's free starter plan to generate and customize a character, highlighting features like prompt building and negative prompts. They also discuss the importance of blending models and using reference images for specific poses, showcasing the potential for creating unique characters while maintaining consistency across different styles and viewpoints.

Takeaways

  • πŸ” AI tools for consistent characters in imagery have been challenging to develop, but a new solution is emerging.
  • 🎨 The AI image generator 'Scenario' has roots in creating assets for games and offers a potential solution for consistent characters.
  • πŸ†“ Scenario offers a free starter plan that is generous enough for extensive testing and basic usage.
  • 🌐 Users can select different aesthetic styles and blend them to create a custom character.
  • 🎭 The interface includes sliders to control the influence of each model on the output, and a prompt box for character description.
  • πŸ“Έ Reference images are crucial for character creation, and users can upload their own or use provided templates.
  • πŸ”„ Iteration is key in the process, as AI will still produce results that need refinement and coherence across different views.
  • πŸ–ΌοΈ Upscaling and enhancing images can be done within Scenario or through external tools like 'Magnific' for better results.
  • πŸ€– Training your own model with multiple views of the character helps the AI learn and produce more consistent results.
  • 🎨 Negative prompts are important to refine the AI's output and avoid unwanted features or styles.
  • πŸ‘Ύ The platform's ability to blend models and allow testing of different combinations leads to a variety of stylistic outcomes.

Q & A

  • What is the main challenge in AI imagery discussed in the script?

    -The main challenge discussed in the script is achieving consistent characters in AI imagery, which has been likened to a Holy Grail due to the difficulty in obtaining satisfactory results.

  • What is Scenario and how does it relate to the discussion?

    -Scenario is a new AI image generator that has its roots in creating AI assets for games. It is presented as a potential solution to the challenge of creating consistent characters in AI imagery.

  • How does the free starter plan in Scenario compare to the paid tiers?

    -The free starter plan in Scenario is quite generous, allowing users to perform a significant amount of testing and creation without any cost. Paid tiers offer additional features and capabilities.

  • What are the aesthetic styles available in Scenario?

    -Scenario offers a number of different aesthetic styles, which users can select based on the vibe they are looking for. These styles can be blended and remixed to achieve desired outputs.

  • How does the prompt system work in Scenario?

    -Scenario has a prompt system that includes a prompt builder and a prompt spark feature. The prompt builder provides keywords or ideas, while prompt spark generates different suggested prompts based on a basic input.

  • What is the importance of negative prompts in Scenario?

    -Negative prompts play a crucial role in Scenario as they help guide the AI to avoid generating certain elements that are not desired, such as elf ears in the example provided.

  • How does the character creation process begin in Scenario?

    -The character creation process begins by selecting a template of either a male or female character, then setting the mode to image to image, and starting to prompt with specific but general characteristics of the desired character.

  • What is the purpose of the upscaling feature in Scenario?

    -The upscaling feature in Scenario is used to enhance the quality of the generated images. It offers options like face restoration, photo optimization, and enhanced upscaling to improve the final output.

  • How can users train their own model in Scenario?

    -Users can train their own model by dragging in their images of the character from different viewpoints and providing descriptive information about the character to give the model a better understanding of the character's identity.

  • What is the role of blending models in Scenario?

    -Blending models in Scenario allows users to combine their own model with other models to achieve different styles and improve the character's consistency across various outputs.

  • What additional tools does Scenario offer for image editing?

    -Scenario offers a canvas function for basic image editing, such as painting out unwanted elements, and a layers tab for trying out different generations of the image.

Outlines

00:00

🎨 Introducing Scenario: AI Image Generator

The paragraph introduces Scenario, an AI image generator with roots in creating assets for games. The speaker acknowledges previous attempts at creating consistent characters in AI imagery and expresses excitement about Scenario's potential. The video will explore a workflow discovered in a thread by Vlados AI, a customer success specialist and AI artist at Scenario. The speaker notes that Scenario offers a free starter plan, which is sufficient for the testing conducted in the video. The focus is on creating a custom character and using the platform's features, such as aesthetic styles and sliders to blend and remix elements. The process involves selecting a model, generating an image, and refining the character through iteration and the use of prompts for different viewpoints.

05:02

🌟 Character Development and Training

This paragraph delves into the character development process within Scenario. The speaker explains how to upscale the character image using the platform's enhanced upscale feature and then edit the image in Photoshop or a similar editor. The character is then prepared for training by separating the front, side, and back views. The speaker emphasizes the importance of negative prompts and provides tips on how to train the model effectively. The goal is to create a consistent character that can be used in various styles and poses without losing its identity. The paragraph also touches on the blending of models to achieve desired results and the use of control nets for specific poses.

10:02

πŸ–ŒοΈ Refining and Experimenting with Scenario

The speaker continues to explore Scenario's capabilities, discussing the use of negative prompts to correct unwanted features like elf ears. The paragraph covers the use of image referencing and control nets for achieving specific poses and refining the character's appearance. The speaker shares their experience with blending models and the importance of iteration in achieving satisfactory results. The paragraph concludes with a mention of Scenario's canvas function for editing images and the potential for future development of the platform. The speaker encourages viewers to explore Scenario for free and provides a link to a PDF guide on creating consistent characters.

Mindmap

Keywords

consistent characters

Consistent characters refer to the creation of characters in AI-generated imagery that maintain a uniform appearance and attributes across different images. This is significant in the context of AI art as it ensures that the character's design remains recognizable and true to its original concept, despite variations in style or viewpoint. In the video, the speaker discusses the challenges and solutions related to achieving consistent characters using an AI tool called Scenario.

AI imagery

AI imagery involves the use of artificial intelligence to generate visual content, such as images or animations. In the video, the term is used to describe the process of creating digital art with the help of AI, particularly in the context of generating characters with consistent features. The speaker highlights the difficulties in achieving high-quality, consistent AI imagery and introduces a tool that aims to address these issues.

Scenario

Scenario is an AI image generator mentioned in the video that is designed to create AI assets for games but has also been adapted for generating consistent characters in AI imagery. The speaker discusses the features and workflow of Scenario, emphasizing its ability to blend different aesthetic styles and create custom characters through a combination of prompts and image references.

aesthetic styles

Aesthetic styles refer to the visual design and artistic approach used in creating images or characters. In the context of the video, the speaker selects different aesthetic styles within the Scenario tool to influence the final look of the generated characters. These styles range from 'Illustrated anime' to 'Creepy Nordic fairy tale,' and they can be mixed and matched to achieve a desired outcome.

prompts

Prompts are the input text or phrases that guide the AI in generating specific images or characters. In the video, the speaker uses prompts to provide direction to the Scenario tool when creating a custom character, including details like age, hair color, and clothing type. The use of prompts is crucial in achieving consistent characters, as they help the AI understand the desired attributes and appearance of the character.

negative prompts

Negative prompts are a feature in Scenario that allows users to specify what they do not want to appear in the generated images. This is important for refining the AI's output and ensuring that the final character design is coherent and free from unwanted elements. In the video, the speaker uses negative prompts to avoid certain features, such as elf ears, that were inadvertently included in the character design.

upscale

Upscale refers to the process of increasing the resolution or quality of an image. In the video, the speaker uses the upscale feature within Scenario to enhance the quality of the generated character images. Additionally, the speaker mentions using an external tool called 'Magnific' for further creative upscaling, which results in higher quality and more detailed images.

training models

Training models in the context of the video involves using a set of images to teach the AI to recognize and generate a specific character. The speaker guides the viewer through the process of training a model with various views of the character to reinforce its features and ensure consistency across different images. This process is essential for creating a custom AI model that can generate images of the character in various contexts and styles.

compose with precision

Compose with precision is a feature in Scenario that allows users to blend their trained models with other existing models to create a more refined and stylized version of their character. By adjusting sliders to control the influence of each model, users can achieve a desired aesthetic while maintaining the consistency of their character's appearance. The speaker demonstrates this by combining the Siri model with other styles, such as 'Creepy Nordic fairy tale,' to produce varied and visually interesting results.

control net

Control net is a tool within Scenario that provides additional options for manipulating the pose and appearance of the generated character. It offers a secondary set of controls that allow users to adjust specific aspects of the image, such as character, landscape, structure, and pose. The speaker uses control net to fine-tune the character's pose, ensuring that the AI's output aligns more closely with the desired outcome.

Highlights

A new AI tool for creating consistent characters in AI imagery has been introduced, addressing a long-standing challenge in the field.

The AI tool, Scenario, is rooted in creating AI assets for games and offers a free starter plan for users to explore its capabilities.

Scenario allows users to blend and remix aesthetic styles to create unique characters, providing a variety of models to choose from.

The interface includes sliders to control the influence of each model on the output, enabling fine-tuning of the character's appearance.

A prompt builder and prompt spark features assist users in generating ideas and refining their character prompts.

Negative prompts play a crucial role in Scenario, helping to shape the character's appearance by excluding undesired features.

Reference images are used to begin the character creation process, with templates provided for ease of use.

The process involves iterating and experimenting with prompts to achieve a coherent character representation across different viewpoints.

Upscaling options are available within Scenario to enhance the quality of the character images.

Photoshop or other photo editors can be used to further refine and prepare the character images for use.

Training your own model with Scenario allows for the creation of personalized characters with specific attributes and poses.

Blending your model with other models can introduce new styles and elements, maintaining character consistency across different aesthetics.

Control net offers advanced options for pose and structure control, allowing for more precise character representation.

Scenario's canvas and layers features enable users to make adjustments and experiment with different generations of the character.

The platform's flexibility and innovative approach to character creation make it an exciting tool for artists and designers.

Creator Tim provides a PDF guide on creating consistent characters with Scenario, offering additional support and resources for users.