How To Create Multiple AI Characters In One Scene

Playground AI
5 May 202410:22

TLDRIn this tutorial, the presenter demonstrates how to create a scene with multiple AI characters without merging their traits. The process involves using a tool like 'canvas' to compose the scene with a background and three characters, removing the background from each character, and then adjusting their size and position to fit the scene. To address inconsistent lighting, an image-to-image transformation is applied with a sunset lighting effect. The characters' details are then manually upscaled by cropping and regenerating them with an image-to-image tool, ensuring better facial and hand details. Finally, the characters are composed back into the scene, with attention given to aligning and blending the details for a seamless final image. The presenter also mentions the development of a new UI called 'playground V3' that will combine canvas and board functionalities.


  • ๐ŸŽจ **Scene Composition**: The process begins by composing a scene with a background and multiple characters to avoid character merging.
  • ๐Ÿงฉ **Asset Management**: Keep original character assets intact to allow for re-editing if needed.
  • ๐Ÿšซ **Background Removal**: Remove the background from each character to facilitate clean integration into the scene.
  • ๐Ÿ” **Detailing**: Ensure clear separation of characters and erase any remnants for a clean look.
  • ๐ŸŒ„ **Lighting Consistency**: Address different lighting on subjects compared to the background by using image-to-image adjustments.
  • ๐Ÿ”„ **Duplication & Adjustment**: Duplicate the scene and manually adjust character sizes and positions for a balanced composition.
  • ๐Ÿ“ **Cropping for Focus**: Crop individual characters tightly, being careful to retain important details like shadows.
  • ๐Ÿ”ง **Manual Upscaling**: Perform manual upscaling to improve details without affecting the overall composition.
  • ๐Ÿ”„ **Iterative Refinement**: Use a combination of cropping, image-to-image adjustments, and in-painting to refine the final image.
  • โš™๏ธ **Tool Utilization**: Utilize tools like canvas and image-to-image to control the scene and achieve desired results.
  • ๐Ÿ“ **Continuous Learning**: Stay updated with new tools and UI changes, such as the upcoming Playground V3, to maintain efficiency in the creative process.

Q & A

  • What is the main limitation mentioned when trying to generate two AI characters in one scene?

    -The main limitation is that the characters may merge and share traits of each other's costume, resulting in an unrealistic scene.

  • What is the first step in creating multiple AI characters in one scene?

    -The first step is to compose the scene with different elements, including a background and individual characters.

  • Why is it advised to keep the original character images when starting the process?

    -It is advised to keep the originals in case things go wrong or if you need to start over to avoid losing progress.

  • How does one remove the background from each character in the scene?

    -One can select each image and use tools like 'alt click and drag' to remove the background, and erase any remnants if necessary.

  • What is the purpose of duplicating the scene and increasing the size manually?

    -This is done to bring the characters into the scene and to control their placement, ensuring they are in the desired positions.

  • Why is the lighting adjustment important when composing the scene with different subjects?

    -Lighting adjustment is important because it ensures that the lighting is consistent across all subjects, creating a more realistic and cohesive scene.

  • What is the recommended image strength for image to image transformation?

    -The recommended image strength for image to image transformation is between 45 to 50, which can be adjusted based on the desired outcome.

  • How can one manually upscale individual characters to improve details without affecting the overall composition?

    -One can manually upscale by cropping each character tightly, increasing the size, and using image to image transformation to pull more details from the image.

  • What is the purpose of regenerating the image with different prompts when doing an image to image upscale?

    -Regenerating the image with different prompts helps to condition the AI to focus on specific subjects and details, improving the final result.

  • How can one ensure that the upscaled image maintains proper details and aligns well with the original composition?

    -By using a template from the original composition, cropping and regenerating each character individually, and then aligning them to match the original composition.

  • What is the advantage of using the manual upscaling method described in the script?

    -The advantage is that it allows for control over the original composition, improvement of details, and minimal impact on other elements of the scene.

  • What is the future update mentioned that will potentially change the user interface for creating AI characters?

    -The future update is Playground V3, which will combine the features of Canvas and Board into a new user interface.



๐ŸŽจ Creating a Scene with Multiple Characters

The video begins with an introduction to creating a scene with multiple characters using a tool called 'canvas'. The speaker explains the limitations of AI models when trying to generate images of two characters together, which often results in a merged and inaccurate representation. To overcome this, the use of canvas is suggested to compose a scene with a background and three separate characters, which can then be combined to create a final image. The process involves removing the background from each character, adjusting their size and position within the scene, and using the 'image to image' feature to apply consistent lighting across all elements. The goal is to achieve a well-lit and cohesive scene that accurately represents the desired characters and setting.


๐Ÿ–Œ๏ธ Manual Upscaling and Detail Tweaking

The second paragraph delves into the process of manually upscaling and tweaking the details of each character to improve the overall image quality. The speaker demonstrates how to crop each character tightly, keeping essential elements like shadows, and then use the 'image to image' tool to upscale the cropped images while maintaining the aspect ratio. This method allows for better facial details and hand clarity. The speaker emphasizes the importance of including the subject in the prompt during the upscaling process to ensure the generated images align with the desired characters. After selecting the best images, the characters are reintegrated into the scene, with attention given to aligning them properly and addressing any inconsistencies in the background. The process, while time-consuming, offers control over the final composition and ensures that the characters and scene elements are detailed and well-defined.


๐Ÿ“ˆ Upcoming Changes and Utilizing Drawings

In the final paragraph, the speaker discusses upcoming changes to the 'playground' tool, hinting at a new user interface that combines features of canvas and board. They assure viewers that despite these changes, the core concepts and tools will remain applicable. Additionally, the speaker mentions another method of controlling and conditioning images through the use of drawings, which is covered in other videos. The video concludes with a reminder of the community engagement and the continuous development of the platform to improve user experience and creative output.



๐Ÿ’กAI Characters

AI Characters refer to artificial intelligence-driven entities that can be programmed to exhibit human-like traits and behaviors. In the context of the video, the creator is discussing the process of generating multiple AI characters within a single scene, which is a task that involves careful composition and manipulation of digital assets to ensure each character is distinct and well-defined.

๐Ÿ’กScene Composition

Scene Composition is the art of arranging visual elements within a frame to create a coherent and aesthetically pleasing image. In the video, the creator focuses on how to effectively compose a scene with a background and multiple AI characters, ensuring that they are not merged and maintain their individual traits.

๐Ÿ’กBackground Removal

Background Removal is the process of digitally separating a subject from its background, typically for the purpose of placing it against a different backdrop or for isolation in a composition. The script mentions removing the background from each character to prepare them for the scene composition.

๐Ÿ’กImage to Image

Image to Image is a technique used in digital art and design where one image is used as a reference or base to create or modify another image. In the video, this technique is employed to adjust the lighting of the characters to match that of the background, creating a more cohesive scene.

๐Ÿ’กLighting Consistency

Lighting Consistency refers to the uniformity of lighting across different elements in a visual scene. It is crucial for creating realism and depth. The video demonstrates how to achieve consistent lighting across the characters and the background using the Image to Image feature.

๐Ÿ’กManual Upscaling

Manual Upscaling is the process of enlarging a digital image while manually enhancing its details to maintain or improve its quality. The video script describes a method of manually upscaling individual character images to correct details and maintain the original composition's integrity.


In-Painting is a digital image editing technique used to fill in or smooth out selected areas of an image. In the context of the video, in-painting is used to seamlessly blend the upscaled characters into the background, removing any visible seams or discrepancies.


Canvas, in the context of the video, refers to a digital workspace where images and characters are composed and manipulated. The creator encourages users to be comfortable with the canvas tool before proceeding with the complex process of creating a scene with multiple AI characters.

๐Ÿ’กPlayground V3

Playground V3 is a new user interface and system that is mentioned as being in development. It is intended to combine the functionalities of 'canvas' and 'board', which are presumably tools used in the creation and composition of AI characters and scenes.

๐Ÿ’กControl Net

Control Net is a term that likely refers to a system or feature within the digital workspace that allows for precise control over the elements of a scene, such as the positioning and attributes of AI characters. The video suggests that utilizing Control Net can help achieve desired results in scene composition.


Drawings, as mentioned in the video, can be used as another method to condition and control images, suggesting that in addition to digital manipulation, traditional artistic skills can also play a role in the creation and refinement of AI character scenes.


The tutorial demonstrates how to create multiple AI characters in one scene using a canvas tool.

The technique avoids merging of characters by composing the scene with separate elements.

Background and three characters are used to create a final scene, preventing character traits from blending.

The process includes removing the background from each character to ensure clear separation.

Duplicating the scene and manually adjusting the size of characters to fit the composition.

Using the left and right bracket keys to bring characters to the front in the scene.

Cropping the background to make it tighter and adjusting character sizes for a better composition.

Addressing lighting inconsistencies between subjects and background by using an image-to-image process.

Selecting an image strength between 45 to 50 for the image-to-image process as a good starting point.

Adding a sunset lighting effect to unify the lighting across all characters in the scene.

Manually tweaking details like hands after the initial image-to-image process.

Performing a manual upscale by duplicating the image and increasing its size for more details.

Cropping individual characters tightly to maintain shadows and avoid changing background details too much.

Using the image-to-image feature with a strength of 45 to generate more detailed images.

Changing the prompt during image regeneration to include the subject for better results.

Lining up the new, upscaled assets and adjusting for any blurry or mismatched areas.

Utilizing the in-painting tool to seamlessly blend regenerated areas with the original background.

The process, while time-consuming, allows for control over the scene and character placement.

An announcement about the development of Playground V3, which will combine canvas and board functionalities.

Mention of using drawings as another method to condition and control images in the canvas tool.