21 Jan 202304:38

TLDRIn this informative video, Binks introduces viewers to the process of AI upscaling using the Stable Diffusion Web UI. After expressing gratitude for the support received on previous videos, Binks guides the audience through the 'Extras' tab to demonstrate how to upscale an image, specifically a painting of his dog, by four times its original resolution. He recommends the 'Swin IR' method for this task, cautioning against using it for extreme upscaling. Binks also discusses the importance of 'gift' and 'visibility' for restoring faces in images and provides a brief guide on changing an image's resolution without cropping. The video concludes with Binks highlighting the potential commercial applications of AI upscaling and encouraging viewers to engage with questions and feedback.


  • 🎨 **AI Upscaling Introduction**: Binks discusses the process of AI upscaling and its various applications in the video.
  • 🐕 **Image Selection**: Binks chooses a painting of his dog to demonstrate the upscaling process, aiming to increase its resolution by four times.
  • 🔍 **Upscaler Options**: Several upscaling methods are mentioned, including Esser, RSR, and Swin IR, with Swin IR being Binks' preferred choice for moderate upscaling.
  • 🚫 **Scaling Limitations**: Binks advises against using Swin IR for very large upscales, such as directly to the maximum of eight times.
  • 🧍 **Face Restoration**: The importance of 'gift' and 'visibility' for faces is highlighted, which helps restore facial details more accurately.
  • 🔧 **Alternative Tools**: 'Code former' is mentioned as an alternative for those who do not have 'gift Gan' installed.
  • ⌛ **Process Monitoring**: Binks notes the absence of a progress display in the Stable Diffusion Web UI but suggests monitoring the web console for updates.
  • 🖼️ **Upscaled Result**: The demonstration results in an upscaled painting that Binks is pleased with and considers hanging on his wall.
  • 💡 **AI Business Idea**: Binks suggests the potential business opportunity of upscaling images for clients as a service, leveraging AI technology.
  • 📊 **Resolution Adjustment**: The video covers how to change an image's resolution, emphasizing the difference between downscaling and cropping to fit.
  • 🔄 **Non-Generative Process**: Binks clarifies that upscaling is not a generative process and does not create additional pixels or content.
  • ✂️ **Cropping for Resolution**: When 'crop to fit' is checked, the image is cropped to match the desired resolution, rather than generating new content.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is about mastering AI upscaling using the Stable Diffusion Web UI, focusing on how to upscale images and the different options available within the tool.

  • Who is the presenter of the video?

    -The presenter of the video is Binks.

  • What does Binks suggest as a good shortcut for upscaling an image?

    -Binks suggests using 'Swin IR four times' as a good shortcut for upscaling an image.

  • What is the maximum recommended upscale factor for Swin IR according to Binks?

    -Binks does not recommend using Swin IR for a really big upscale, such as going straight to the maximum slider of eight times.

  • What are the two features that Binks mentions are important for upscaling faces?

    -The two features Binks mentions for upscaling faces are 'gift' and 'visibility', which help restore faces more true to their original form.

  • How does Binks demonstrate the upscaling process?

    -Binks demonstrates the upscaling process by choosing a painting of his dog and upscaling it by four times using the Stable Diffusion Web UI.

  • What is the potential business application of AI upscaling that Binks mentions?

    -Binks suggests that there could be a market for upscaling people's images for a flat rate using AI, which could be a business opportunity for someone savvy in the field.

  • What is the difference between upscaling and scaling to a certain resolution?

    -Upscaling involves increasing the resolution of an image and potentially generating new pixels to fill in the details. Scaling to a certain resolution, on the other hand, simply changes the dimensions of the image without generating new content, which may involve cropping if 'crop to fit' is checked.

  • Why is it important to remember that upscaling is not a generative process?

    -It's important to remember that upscaling is not a generative process because it does not create extra pixels or new content; it only increases the size of the existing pixels, which can lead to a loss in quality if not done properly.

  • How can one follow Binks for more content?

    -To follow Binks for more content, viewers are encouraged to like, subscribe, and join the Discord server, the link to which is provided in the video description.

  • What is Binks' stance on the potential of AI in art and other practical applications?

    -Binks believes that AI is not just about art but can also be extremely useful in practical applications, such as upscaling images for various purposes.

  • What does Binks suggest for those who want to learn more about the different upscaling options?

    -Binks suggests that viewers can do more research on the different upscaling options and their acronyms on their own to understand their specific uses and benefits.



🎨 Introduction and Upscaling Overview

In this video, Binks expresses gratitude for the support received on previous videos and encourages viewers to like, subscribe, and join the Discord server for future content. The main focus of the video is on image upscaling using AI and stable diffusion technology. Binks demonstrates how to upscale an image of his dog by four times its original resolution using different upscaling methods, emphasizing 'Swin IR' as a favorite option. He also discusses the importance of the 'gift' and 'visibility' features for restoring faces in upscaled images and mentions the potential business applications of AI upscaling services.



💡AI Upscaling

AI Upscaling refers to the process of using artificial intelligence to increase the resolution of an image or video. In the video, Binks discusses using AI and stable diffusion techniques to upscale an image of his dog by four times its original resolution, which is a key focus of the content.

💡Stable Diffusion

Stable Diffusion is a term that likely refers to a specific AI model or software used for image processing, including upscaling. Binks mentions 'stable diffusion stuff' as part of the ongoing content he plans to provide, indicating its importance in the context of the video.


To upscale in the context of the video means to enhance the size and potentially the quality of an image without losing its details. Binks chooses to upscale an image of his dog, aiming for a higher resolution for printing and display purposes.


Resolution is the number of pixels in an image or display, which determines its clarity and detail. Binks talks about increasing the resolution of his dog's painting to four times its base resolution, which is a central part of the upscaling process he demonstrates.

💡Extras Tab

The Extras Tab is a feature within the software interface that Binks uses to access the upscaling tools. It is where users can find additional options for image manipulation, including upscaling, which is the primary topic of the video.


Shortcuts in this context refer to quick methods or commands that Binks suggests for using the upscaling feature. He mentions 'Esser', 'rsr', and 'swin IR' as shortcuts for different upscaling options, which are essential for viewers to follow along with the tutorial.


GAN stands for Generative Adversarial Network, a type of AI algorithm used in the upscaling process to restore and enhance image details, particularly faces. Binks discusses the importance of GAN for maintaining the integrity of faces in upscaled images.


In the context of the video, 'generate' is the action of creating or producing an upscaled image using the AI upscaling tools. Binks instructs viewers to click 'generate' to start the upscaling process for their chosen image.


The console, in this case, refers to the web console, a tool that developers use to interact with the browser and debug web applications. Binks suggests looking at the web console to monitor the progress of the upscaling process, as it displays the actions and progress percentage.

💡Crop to Fit

Crop to Fit is an option that allows the software to adjust the image dimensions by cropping out parts of the image to fit the desired resolution. Binks demonstrates the effect of using and not using this option when changing the width and height of the image.

💡Discord Server

A Discord server is a chat community on the Discord platform where people with similar interests can communicate. Binks mentions having a Discord server and provides a link in the video description for viewers to join and engage with the community.


Introduction of the topic and appreciation for viewer support.

Mention of upcoming content related to AI and stable diffusion.

Overview of the video focus—upscaling images using a web UI.

Process of selecting an image for upscaling, using a personal painting as an example.

Explanation of different upscaling options available.

Recommendation of specific upscaling tools and their appropriate uses.

Tips on using the upscaler for best results, especially on images with faces.

Step-by-step guide on generating an upscaled image.

Lack of a progress display feature in the stable diffusion web UI.

Personal anecdote on the value of upscaling a painting for printing.

Discussion on the potential business opportunities in AI-based image upscaling.

Additional features of the web UI, like adjusting image resolution.

Explanation of the 'crop to fit' option and its effects on the image.

Clarification that the process isn't generative and what that means for image quality.

Closing remarks, inviting viewers to comment, like, and subscribe for more content.