An AI artist explains his workflow

Vox
2 May 202308:18

TLDRThe video script introduces Stelfie, a humorous and clumsy character who embarks on time-traveling adventures. The creator uses Stable Diffusion and artistry to depict Stelfie in a boxing match with Muhammad Ali. The process involves sketching, experimenting with prompts, and refining the image in Photoshop and Procreate. The artist emphasizes the importance of controlling the AI, rather than being controlled by it, and sees the collaboration as an opportunity for new artistic expression.

Takeaways

  • 🎨 The artist uses Stelfie as an alter ego character in their creative projects, combining Stable Diffusion with traditional art skills.
  • ⏱️ The initial goal was to create a scene featuring Stelfie in a boxing match with Muhammad Ali, showcasing the potential of AI in art.
  • 📝 The creative process starts with a sketch, followed by experimenting with various prompts to find a suitable initial pose.
  • 🖌️ Photoshop is utilized to recreate poses and refine details that AI might not capture accurately, such as facial expressions and body shapes.
  • 🔄 ControlNet, a tool for Stable Diffusion, is mentioned as potentially saving significant time if recreating a past pose.
  • 🧪 Different samplers like Euler and DPM are used in the process, each affecting the realism and details of the artwork, particularly for textures like skin.
  • 🔢 Parameters such as 'steps', 'inpaint', and 'outpaint' are crucial for guiding Stable Diffusion in refining parts of the image or imagining new ones.
  • 🤖 AI is used for about 50% of the work, with the remaining 40% done in Photoshop and 10% in Procreate, highlighting a hybrid approach to digital art.
  • 👤 The artist emphasizes the importance of driving the AI rather than being driven by it, seeing it as an opportunity for new creative avenues.
  • 🖼️ The final artwork is a collaboration between the artist and AI, with the artist's traditional skills playing a significant role in achieving a realistic and creative outcome.

Q & A

  • Who is Stelfie and what is his character like?

    -Stelfie is a humorous and clumsy character who engages in time travel and has incredible adventures. He is also an alter ego of the speaker, though they physically differ.

  • What was the original purpose of starting the Stelfie project?

    -The original purpose was to showcase the potential of Stable Diffusion combined with good artist skills, specifically aiming to capture a scene where Stelfie engages in a boxing match with Muhammad Ali.

  • How does the speaker typically begin creating a scene with Stable Diffusion?

    -The speaker usually starts by drawing a sketch to outline the scene they wish to create.

  • What challenges does the speaker mention about using Stable Diffusion and similar models?

    -The speaker mentions that these models can be cheeky and might lead one away from the original idea due to their tendency to produce unexpected results.

  • How does the speaker refine the initial pose in the artwork?

    -If a suitable pose is not found through Stable Diffusion, the speaker uses Photoshop to recreate the desired pose manually.

  • What role does ControlNet play in the process?

    -ControlNet is an extension that, if used today, would significantly reduce the time required to reproduce a pose that took much longer to create in the past.

  • What is the importance of samplers in the artwork creation process?

    -Samplers are crucial for achieving realism and detail in the artwork, such as accurately replicating skin texture.

  • How does the speaker balance the use of Stable Diffusion, Photoshop, and Procreate in the creation process?

    -The speaker estimates that 50% of the work is done in Stable Diffusion, 40% in Photoshop, and 10% in Procreate.

  • How did the speaker train the model specifically for Stelfie's face?

    -The speaker created Stelfie in 3D, took snapshots of his face from various angles, and used those images to train the model, saving a keyword for future use.

  • What is the significance of noise strength in Stable Diffusion?

    -Noise strength is important as it provides control over the image, helping to achieve better results, especially for faces.

  • How did the speaker modify the depiction of Muhammad Ali in the artwork?

    -The speaker asked Stable Diffusion to create a face like Muhammad Ali and then manually adjusted features such as the nose, jaw, and eyes in Photoshop to achieve a more realistic representation.

  • What does the speaker suggest about the role of the artist in the age of AI?

    -The speaker views the overall process as a joint effort with AI, seeing it as an opportunity for new talent to explore a different branch of art and open up new ways of being creative.

  • Why does the speaker use their own hands in the artwork?

    -The speaker uses their own hands in the artwork because reproducing hands is extremely challenging, so they take pictures of their hands, clean them up, and paste them onto the artwork.

Outlines

00:00

🎨 Creative Process with AI and Art

The paragraph discusses the creative journey of an artist using Stable Diffusion and other tools to bring their character, Stelfie, to life in a boxing match with Muhammad Ali. The artist begins with a sketch and uses Stable Diffusion to generate ideas, but finds the model to be cheeky and potentially distracting from the original vision. They then utilize Photoshop to refine the pose and details, highlighting the importance of samplers for realism. The artist also discusses the use of ControlNet, steps, inpaint, and outpaint features in refining the artwork. The process involves a combination of Stable Diffusion, Photoshop, and Procreate, with a focus on creating a realistic and unique Stelfie character, not super fit, and a realistic Muhammad Ali. The artist emphasizes the importance of the artist's role in guiding the AI and sees the collaboration as an opportunity for new talent in the art world.

05:02

🖌️ Refining Art with Traditional and Digital Techniques

The artist delves into the challenges of capturing Muhammad Ali's likeness and the physicality of the scene. They discuss the need for manual adjustments in Photoshop, such as cropping, warping, and painting to achieve the desired look. The artist reflects on their understanding of Ali's appearance and the efforts to make it realistic, including adjusting muscle definition. The narrative also touches on the artist's background in traditional and digital art, emphasizing the artist's role in driving the creative process with AI as a tool. The artist's hands-on approach to creating difficult elements like hands showcases the blend of traditional artistry and digital innovation.

Mindmap

Keywords

💡Stable Diffusion

Stable Diffusion is an AI model that generates images from textual descriptions. It is a form of artificial intelligence that uses machine learning to understand and create visual content based on the input it receives. In the context of the video, Stable Diffusion is used to create the initial sketches and visual concepts for the character Stelfie and his adventures, showcasing the potential of AI in the artistic process.

💡Artist Skills

Artist skills refer to the technical abilities and creative talents that an individual uses to produce visual art. These skills can include drawing, painting, and understanding composition, color theory, and other elements of visual design. In the video, the artist combines their artist skills with Stable Diffusion to create a unique character and scene, emphasizing the importance of human creativity alongside AI technology.

💡Alter Ego

An alter ego is a second self or a different aspect of a person's personality that they may present to the world. In the context of the video, Stelfie is described as an alter ego of the artist, although physically they are completely different. This concept is used to explore different personas and characteristics in the artistic creation process.

💡Photoshop

Photoshop is a widely used software program for image editing and manipulation. It provides a suite of tools that allow artists to alter and enhance digital images. In the video, the artist uses Photoshop to refine the poses and details of the characters, demonstrating the software's role in the digital art creation process.

💡ControlNet

ControlNet is an extension or tool that aids in the reproduction of specific visual elements in artwork. It can streamline the process of recreating a particular pose or detail by providing a framework or guide for the artist. In the context of the video, ControlNet is mentioned as a tool that could significantly reduce the time required to reproduce a pose that was previously created.

💡Samplers

Samplers in the context of AI image generation refer to different algorithms or methods used to generate or 'sample' the final output based on the input data. Each sampler can affect the level of realism and detail in the generated images. For instance, the Euler sampler is described as synthetic and fake, while DPM is noted for working well on replicating skin textures.

💡Steps

In AI image generation, 'steps' refer to the number of iterations the AI performs to refine and improve the image based on the input prompt. A higher number of steps can lead to more detailed and refined outputs, but it may also take longer to process. The artist in the video discusses the strategic use of steps to balance the level of detail and the time taken to generate the image.

💡Inpaint and Outpaint

Inpaint and outpaint are features in AI image generation that allow for the modification of existing images and the creation of new content based on the existing data. 'Inpaint' involves changing specific parts of an image, while 'outpaint' involves generating content that extends beyond the original boundaries of the image, based on the AI's understanding of the content within.

💡3D Modeling

3D modeling is the process of creating a three-dimensional representation of an object or character using computer software. This technique allows for the visualization of the subject from various angles and can be used to create a more accurate and detailed understanding of the form and appearance of the subject. In the video, the artist used 3D modeling to capture Stelfie's face from different angles, which was then used to train a specific model for generating his likeness.

💡Noise Strength

Noise strength in AI image generation refers to the level of randomness or variation introduced into the generated image. Adjusting noise strength can affect the overall quality and detail of the image, with higher values potentially leading to more detailed but also more unpredictable results. The artist in the video discusses noise strength as an important parameter for controlling the final appearance of the generated images, especially when dealing with faces.

💡Digital Art

Digital art is a form of artistic expression that uses digital technology as a primary tool for creating and manipulating images. This can include a wide range of techniques, from digital painting and graphic design to 3D modeling and AI-generated art. In the video, the artist discusses their transition from traditional art to digital art, highlighting the opportunities that new technologies provide for artists to explore new forms of creativity.

Highlights

Stelfie is a character that embodies humor and clumsiness while engaging in time-traveling adventures.

The creator views Stelfie as an alter ego, despite their physical differences.

The project was initiated to showcase the potential of Stable Diffusion combined with artistic skills.

The creative process begins with a sketch to capture the intended scene.

Diffusion models like Stable Diffusion can be cheeky and may deviate from the original idea.

Random prompts are used to find a good initial pose for the artwork.

Photoshop is utilized to recreate poses not found with Stable Diffusion.

ControlNet could significantly reduce the time needed to reproduce a pose with Stable Diffusion.

Samplers play a crucial role in achieving realism and detail in the artwork.

Parameters like steps, inpaint, and outpaint are essential in refining the artwork.

The process involves a combination of Stable Diffusion, Photoshop, and Procreate.

A specific model trained on Stelfie's face is used for his portrayal.

The noise strength parameter in Stable Diffusion allows for control over the image outcome.

Muhammad Ali's face was created using Stable Diffusion and then modified in Photoshop.

The artist aimed for Stelfie to appear not super fit, with a more realistic portrayal.

The artwork involves a significant amount of manual adjustment in Photoshop.

The artist believes in driving the machine rather than being driven by it, emphasizing the importance of the artist's role.

The use of AI in art is seen as an opportunity for a new branch of creativity.

The artist's hands are often used in the artwork due to the challenge of reproducing hands digitally.

Casual Browsing

you're not an artist 🤷‍♂️

2024-04-07 12:40:01

Can an artist spot AI-generated art?

2024-04-03 09:50:01

The Secrets to Work-Life Balance as an Artist in 2024

2024-04-04 19:50:01

I turned ChatGPT into an AI Artist, and now it's better than I am...

2024-03-31 22:40:01

Perturbing His Pastor

2024-04-09 19:05:01

AI vs ARTIST - WHICH IS BETTER?

2024-04-06 03:10:01