【Stable Diffusion】構図・アングル・視線の呪文集プロンプトをまとめて紹介

AIジェネ【AIイラスト生成の情報発信】
26 Aug 202308:23

TLDRThe video script offers a comprehensive guide on creating compositions and selecting angles for AI-generated images. It emphasizes the importance of mastering specific 'spells' or prompts, such as 'head shot' and 'cowboy shot', to achieve desired compositions like full-body views or shots from above or below. The script also discusses image size adjustments and the use of prompts to control the character's line of sight, ultimately aiming to improve the quality and accuracy of AI-generated content without extensive trial and error.

Takeaways

  • 🎨 Understanding composition and angle spells is crucial for efficient image generation without the need for repetitive generation processes.
  • 📸 To capture an image from the chest area upwards, use the 'head shot' composition spell, which often extends to the chest area despite its name.
  • 🗿 Generating a square image size, like 512x512, can facilitate easier and potentially higher quality compositions.
  • 👤 For compositions starting from the thigh area, the 'cowboy shot' spell is effective, though it may initially default to images with a hat, which can be mitigated by adding ((no hat1.3)) to the prompt.
  • 📏 Adjusting the image size to a vertical长方形, such as 512x1024, can improve the success rate of generating thigh-focused compositions without hats.
  • 🚶‍♂️ To depict a full body composition, using 'full body' as a prompt along with an image size of 512x1024 or 512x768 with 'hires.fix' can enhance the quality, especially around the eyes.
  • 👀 The quality of the eyes in the generated image can be a significant factor, thus using 'hires.fix' may increase processing time but dramatically improves eye detail.
  • 🎯 For directional angles, using specific spells like 'shoot from front', 'above', 'below', 'side', and 'behind' allows for precise control over the perspective.
  • 👀 Controlling a person's line of sight in the image can be achieved with spells like 'looking viewer', 'up', 'down', 'side', 'back', and 'looking away'.
  • 🔍 If the desired composition or angle is not achieved, using methods like 'img2img', 'openpose', or changing the model may offer solutions.
  • 📚 The video script serves as a comprehensive guide for users to master the use of composition and angle spells for effective image generation.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is about creating compositions and controlling angles in image generation using various spells or prompts.

  • How can one generate an image from the chest area upwards?

    -To generate an image from the chest area upwards, you can insert the 'head shot' spell in the prompt.

  • Why should certain spells like 'skirt' or 'denim' be removed from the prompt?

    -These spells should be removed because they can interfere with the desired composition, causing the lower body elements to not generate properly.

  • What is the recommended image size for generating high-quality compositions?

    -A square image size, such as 512 by 512, is recommended as it makes it easier to generate high-quality images.

  • How can the tendency to generate a hat in a 'cowboy shot' be reduced?

    -By adding '(no hat 1.3)' after 'cowboy shot' in the prompt, the tendency to generate a hat can be reduced.

  • What is the recommended method for composing a full-body image?

    -The most recommended method for a full-body composition is to set the width to 512 and the height to 768, and use 'hires.fix' for the highest quality.

  • How can the angle of the generated image be controlled?

    -The angle can be controlled by using specific spells like 'shoot from front', 'shoot from above', 'shoot from below', 'shoot from side', and 'shoot from behind'.

  • What spell can be used to make the generated character look at the camera?

    -The 'looking viewer' spell can be used to make the generated character look at the camera.

  • If a desired composition or angle is not achieved, what alternative methods can be tried?

    -If the desired composition or angle is not achieved, one can try changing the image size, using 'img2img', 'openpose', or 'openpose editor' based on a reference image, or changing the model.

  • What is the significance of understanding and inputting spell content?

    -Understanding and inputting spell content allows for more efficient image generation, reducing the time spent until the ideal image is produced, and improving the overall quality of the generated images.

  • What is the role of 'Ai Gene' in the context of the video?

    -Ai Gene is mentioned as a source that disseminates information about generated AI, suggesting that the video content is educational and aimed at enhancing the viewers' knowledge of AI-generated content.

Outlines

00:00

🎨 Understanding Composition and Angle Viewpoints

This paragraph discusses the intricacies of creating visual compositions using specific spells to capture the desired image. It emphasizes the importance of mastering composition and angle viewpoint spells to generate images efficiently without the need for multiple iterations. The paragraph explains how to generate images from different parts of the body, such as chest shots and cowboy shots, and how to adjust the image size for optimal results. It also highlights the use of spells to control the character's line of sight, such as looking at the camera, looking up, down, or away, and the impact of these spells on the final image. The paragraph concludes with a recommendation to use 'img2img' or 'openpose' if the desired composition or angle is not achieved through spells alone.

05:05

🔍 Advanced Techniques for Image Generation

The second paragraph delves into advanced techniques for achieving specific compositions and angles in image generation. It addresses the challenges of generating images with upward or downward gazes and the necessity of using spells like 'looking up' or 'looking down' to achieve these effects. The paragraph also explores the use of 'looking side' and 'looking back' spells to capture images from unique perspectives. It suggests increasing the value after the colon in spells like 'looking away' to ensure the desired outcome. The paragraph further discusses the potential need to adjust image sizes or switch to different models if the initial attempts do not yield satisfactory results. It concludes with a reminder to subscribe to the channel for more content on generated AI and thanks the viewer for their interest.

Mindmap

Keywords

💡Composition

In the context of the video, 'composition' refers to the arrangement of elements within an image, specifically how the character is positioned and portrayed. It is a crucial aspect of creating visually appealing and meaningful artwork. The video provides various 'spells' or techniques to achieve different compositions, such as 'head shot' for focusing on the upper body or 'cowboy shot' for a lower perspective. The term is central to the video's theme as it directly influences the final output of the generated image.

💡Image Size

'image size' pertains to the dimensions of the generated image, which significantly impacts the quality and the ability to capture the desired composition. The video emphasizes the importance of adjusting image size to avoid common issues like eye distortion or incomplete body representation. For instance, a square image size like 512x512 may not be ideal for a full-body image, whereas 512x1024 can improve the quality of a standing composition.

💡Spells

'Spells' in this context are specific terms or prompts used in the image generation process to guide the AI in creating particular effects or capturing certain details. They are essential tools for artists to achieve the desired outcome without having to repeatedly generate images through trial and error. The video discusses various spells such as 'head shot', 'cowboy shot', and 'hires.fix', which directly influence the composition and quality of the generated images.

💡Angles

The term 'angles' refers to the perspective from which the image is captured. In the video, it is discussed how different angles can dramatically change the viewer's perception of the image. The script provides various angle spells such as 'shoot from above', 'shoot from below', and 'looking viewer', which guide the AI in creating images from specific viewpoints. The choice of angle is integral to the video's message as it affects the storytelling and emotional impact of the generated art.

💡Eye Quality

'Eye quality' refers to the clarity, detail, and accuracy with which the eyes of the character are depicted in the image. High-quality eyes can make the image more engaging and realistic. The video emphasizes the importance of image size and specific spells like 'hires.fix' in improving eye quality, which in turn enhances the overall visual appeal of the generated artwork.

💡Full Body

The term 'full body' indicates a complete representation of the character from head to toe. In the video, achieving a high-quality full-body image is a common goal, and various techniques are discussed to accomplish this. The script provides guidance on how to use specific spells and image sizes to generate a full-body composition without common issues like eye distortion.

💡Cowboy Shot

A 'cowboy shot' is a specific composition technique mentioned in the video, which focuses on generating an image from the middle of the thigh area. It is a unique spell that allows artists to capture a lower angle perspective. However, the video also notes that this spell often results in the character wearing a hat, and additional measures like adding '(no hat1.3)' are suggested to avoid this.

💡Hires.Fix

'hires.fix' is a spell or technique highlighted in the video that aims to enhance the quality of the generated image, particularly the eyes. It is part of the effort to improve the overall visual appeal and detail of the artwork. The use of 'hires.fix' is recommended when a higher level of detail and quality is desired, understanding that it may increase the processing time for image generation.

💡Negative Prompt

A 'negative prompt' is a technique used in the image generation process to exclude certain elements from the final output. In the video, it is mentioned as a way to avoid unwanted features like hats in the 'cowboy shot'. By adding a negative value after the unwanted element, the artist can guide the AI to omit that specific detail from the image.

💡Img2img

'img2img' is a method mentioned in the video for generating images based on a reference image. This technique allows artists to input an existing image as a guide for the AI to create a new image with similar composition and style. 'img2img' is a valuable tool when the desired composition or angle cannot be achieved through spells alone, offering an alternative approach to image generation.

💡Openpose

'Openpose' is a technology referenced in the video that can be utilized to generate images with specific poses or postures. It serves as an alternative method when the standard spells and techniques do not yield the desired results. 'Openpose' can be particularly helpful in capturing complex poses or dynamic angles that might be challenging to achieve otherwise.

Highlights

The importance of mastering composition and angle spells for efficient image generation.

Using 'head shot' to generate a composition from the chest area upwards.

The necessity of removing spells like 'skirt' or 'denim' when aiming for a specific composition.

Creating a square image size, such as 512x512, for easier generation.

Generating a high-quality image from the chest to the top without eye collapse.

The 'cowboy shot' for composing from the middle of the thigh.

Preventing the generation of a hat in the 'cowboy shot' by adding ((no hat1.3)) to the prompt.

Adjusting image size to 512x1024 for easier composition from the upper thighs without a hat.

Entering 'full body' to compose the entire body in the image.

Setting the width to 512 and height to 1024 for better quality in a standing composition.

Recommendation to use 'hires.fix' for the highest quality in a full-body image.

Entering 'shoot from front' for a frontal angle with the same line of sight as the character.

Generating an angle from above by entering 'shoot from above'.

Using 'looking viewer' to make the character look at the camera.

Specifying the angle of gaze with 'looking up: 1.2' or 'looking down: 1.2'.

The possibility of using 'img2img', 'openpose', or 'openpose editor' when spells do not yield the desired result.

Changing the model as a last resort if the desired composition or angle is not achieved.

Casual Browsing

【Stable Diffusion】ポーズの呪文集プロンプト一覧を紹介！

2024-04-23 01:10:00

Stable Diffusionで高品質な美少女ができるプロンプト呪文集をまとめて紹介

2024-04-23 01:35:00

【Stable Diffusion】背景を上手く生成させるためのコツと呪文集をまとめて紹介

2024-04-23 00:45:00

Stable Diffusionで手の指を修正するプロンプト呪文やDepth Libraryの使い方について解説

2024-03-31 02:40:00

Stable Diffusionの手の修復方法のまとめ

2024-04-16 23:10:00