BETTER than PROMPTS - The Future of AI Composition

Olivio Sarikas
3 May 202309:22

TLDRIn this video, the presenter, known as the 'king of AI,' guides viewers through the process of creating a visually striking composition using advanced AI techniques. The focus is on achieving more control over elements such as composition, colors, and details. The journey begins with separating the foreground from the background, using selection tools to isolate a female hero with angel wings on a rooftop against a burning city backdrop. The artist then adds wings to the character and adjusts their brightness using curve adjustments. A burning city is placed in the background for contrast. The process continues with exporting the composition at a reduced resolution and refining the character's face using an AI-generated 'beautiful face' prompt. The video also demonstrates combining a portrait with an abstract background and using text-to-image AI to create a 'beautiful African queen' with a colorful abstract background. The presenter shares a trick for using control nets to enhance the composition and concludes with a teaser for a future live stream. The video is an engaging tutorial on leveraging AI for creative image composition.

Takeaways

  • 🎨 **Image Description with Images**: The best way to describe an image is by using another image, giving more control over elements, composition, and colors.
  • ✂️ **Separating Foreground and Background**: Use selection tools to separate the main subject from the background for more creative flexibility.
  • 🖼️ **Adding Angel Wings**: Incorporate angel wings into a composition by copying and pasting them into the desired position and adjusting their brightness.
  • 🌇 **Creating a Burning Cityscape**: Add a dramatic background like a burning city by layering and adjusting the size and position to fit the composition.
  • 🔍 **Refining Selections**: Use refine and mask tools to clean up selections and ensure a smooth transition between elements.
  • 👩 **Character Customization**: Customize the character's appearance, such as changing the face, using specific prompts and tools for fine-tuning.
  • 🖥️ **Software and Tools**: Utilize software like Affinity Photo and tools like control nets for advanced image manipulation and composition.
  • 📐 **Adjusting Brightness and Contrast**: Use curve adjustments to modify the brightness and contrast of elements within the composition.
  • 🧩 **Combining Portraits with Abstract Backgrounds**: Experiment with combining portraits and abstract backgrounds to create unique and artistic images.
  • 📈 **Exporting and Resolution**: When exporting, reduce the resolution to a manageable size without compromising the quality of the image.
  • 🔧 **Technical Settings for Image Generation**: Use specific settings like Euler, steps, restore phase, CFG scale, and denoise strength for generating images.
  • 📚 **Learning from Mistakes**: Even experienced creators make mistakes; it's important to learn from them and continue refining the process.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is about creating a composition using AI, with a focus on having more control over elements like the composition, colors, and overall design.

  • What does the speaker describe as the best way to describe an image?

    -The speaker describes using another image as the best way to describe an image, suggesting a visual reference provides more control and clarity.

  • What is the first step the speaker takes to separate the woman and the foreground from the background in the composition?

    -The first step is to right-click on the layer, rasterize it, then use the selection brush in add mode to select the woman and the rooftop in the foreground.

  • How does the speaker adjust the brightness and contrast of the angel wings in the composition?

    -The speaker selects both wings, groups them, and then uses the Curves adjustment in Affinity Photo to increase the brightness and introduce contrast to the wings.

  • What is the process for adding a burning city background to the composition?

    -The speaker uses the 'File' and 'Place' options to load the burning city image, resizes it as needed, and positions it behind the model and wings in the composition.

  • What is the speaker's approach to exporting the final composition?

    -The speaker reduces the resolution on the long side to about 1600 pixels and then exports the composition.

  • What is the speaker's method for enhancing the face of the character in the composition?

    -The speaker sends the composition to 'Inpaint' to paint out the original face of the character and then uses a new prompt 'beautiful face' to generate an improved face using only mask and denoising.

  • How does the speaker upscale the image with tiles?

    -The speaker uses the 'Image to Image' option, sets the width and height, adjusts the denoise strength, CFG scale, and sampling method to DPM++ 2sa Keras, and enables tile resample control net to upscale the image.

  • What is the trick the speaker uses when combining a portrait with an abstract background?

    -The speaker opens the portrait in Affinity Photo, places the abstract background, and uses the selection brush to paint out the model and mask the background, then adjusts the position and rotation of the background to create a halo effect around the model's head.

  • How does the speaker use text to image to create a composition with an African queen and an abstract background?

    -The speaker inputs a prompt describing the desired composition, uses DPM++ 2sa Keras sampling, and applies two control nets: one for depth and the other for open pose, to generate a composition with a colorful abstract background.

  • What happened to the part of the video where the speaker created an image with a bear holding a red balloon and a girl sitting in his lap?

    -The speaker accidentally deleted that part of the video but mentioned the possibility of recreating it during a live stream.

  • What is the speaker's final call to action for viewers of the video?

    -The speaker encourages viewers to like the video if they enjoyed it and to look forward to future content, suggesting the possibility of a live stream for the deleted part of the video.

Outlines

00:00

🎨 Creative Image Composition with AI

The first paragraph introduces a method for creating compelling compositions using AI, emphasizing the need for more control over elements like composition, colors, and details. The speaker, referring to themselves as the 'king of AI,' guides viewers through the process of creating an image of a female hero with angel wings against a burning city backdrop. The process involves separating the foreground from the background, adding angel wings to the character, adjusting the wing's brightness and contrast, and placing a burning city in the background. The speaker also discusses the technical aspects of exporting the image and using specific settings for the AI to generate the final composition.

05:03

🖼️ Enhancing Images with Text-to-Image Techniques

The second paragraph demonstrates another technique for enhancing images using text-to-image AI. It involves combining a portrait with an abstract ornamental background. The process includes removing the original background, placing the abstract background behind the model, and adjusting it to create a halo effect around the model's head. The speaker then guides viewers on how to export the composition, use text-to-image AI with specific prompts and settings, and further refine the image by inpainting and upscaling. The paragraph concludes with a note about an accidentally deleted part of the video and an invitation to a live stream for a demonstration of creating an image with a bear and a girl.

Mindmap

Keywords

💡AI Composition

AI Composition refers to the process of creating music, art, or other creative works using artificial intelligence. In the video, the term is used to describe the advanced techniques for generating images with AI, emphasizing the control over elements like composition and colors.

💡Rasterize

Rasterize is a term used in graphic design to convert vector graphics into a raster (pixel-based) format. In the script, the speaker uses 'rasterize' when instructing how to transform a layer into a format that allows for further editing, such as making selections.

💡Selection Brush

The Selection Brush is a tool in graphic design software used to select parts of an image. The video demonstrates using the Selection Brush to isolate the woman and the rooftop from the background, which is a crucial step in the AI composition process.

💡Refine

Refine is a process that improves or sharpens an image selection. In the context of the video, 'refine' is used to perfect the selection of the woman and the rooftop, ensuring a clean separation from the background for the composition.

💡Wings

In the video, 'wings' are a visual element added to the female hero character to enhance the image's aesthetic and thematic impact. The wings are manipulated and adjusted for brightness and contrast to fit the desired artistic vision.

💡Burning City

The 'burning city' represents the apocalyptic backdrop against which the female hero stands. It's a key component of the image's narrative, adding a dramatic and intense atmosphere to the composition.

💡Curves Adjustment

Curves Adjustment is a tool in photo editing software that allows for the manipulation of the tonal range of an image. The video uses this tool to adjust the brightness and contrast of the wings, making them stand out more in the composition.

💡Export

Export in the context of graphic design and video production refers to the process of saving or rendering a final version of the work. The script mentions exporting the composition with a specific resolution to prepare it for sharing or further use.

💡CFG Scale

CFG Scale is a parameter used in AI image generation that affects the quality and detail of the output. In the video, it's one of the settings adjusted to achieve the desired level of detail in the AI-generated images.

💡Inpaint

Inpaint is a process used in image editing to fill in or correct parts of an image. The video mentions using 'inpaint' to fix the eyes of the character, which is an essential step for achieving a polished final result.

💡Upscale

Upscale refers to increasing the resolution or size of an image without losing quality. The video script describes using an 'upscale' technique to enhance the final composition, making it larger and more detailed.

Highlights

The future of AI composition is described as being better than using traditional prompts, offering more control over elements like composition and colors.

AI is used to create a female hero standing on a rooftop with angel wings and a burning city in the background.

The process involves separating the woman and the foreground from the background using selection tools.

The use of an image to describe another image is emphasized for better control in AI composition.

Adding angel wings to the female hero using selection and copy-paste techniques.

Adjusting the brightness and contrast of the wings using Curve adjustments in Affinity Photo.

Incorporating a burning city into the background to create a dramatic scene.

Exporting the composition with a reduced resolution for online sharing.

Using a prompt to guide the AI in creating a specific image composition.

The use of Euler, restore phase, and CFG scale in the settings for the AI composition process.

Selecting the most appealing composition from the results generated by the AI.

In-painting and upscaling techniques are used to refine and enhance the final image.

Combining a portrait with an abstract background to create a unique composition.

Using control nets and depth layers to guide the AI in creating detailed and accurate compositions.

The ability to adjust the pre-processor resolution and sampling method for better image quality.

The use of Pixel Perfect and Open Pose control nets for more precise control over the composition.

The final composition is showcased, demonstrating the power and flexibility of AI in image creation.

The video also mentions an accidental deletion of a part where a bear holding a red balloon and a girl sitting in his lap was created.

The presenter offers to recreate the deleted composition in a live stream if viewers are interested.