Tutorial: Making DMT Replications in Midjourney / Magnific
TLDRIn this tutorial, the creator shares a method for generating high-resolution, complex DMT hyperspace AI replications using the Midjourney model version 5, which is considered the best for psychedelic art. The process involves using specific prompts with a focus on words like 'cute' and 'structure' to yield non-evil and non-anthropomorphic entities. The creator emphasizes the importance of rerolling and adjusting settings in both Midjourney and Magnific to achieve the desired intricate, interdimensional structures. The tutorial also covers techniques to minimize visual errors and enhance the final image's clarity. The creator clarifies that the goal is not to produce art for commercial purposes but to replicate psychedelic experiences through AI, treating it as a scientific endeavor.
Takeaways
- 🎨 **Use Midjourney Model Version 5**: The speaker suggests that version 5 of the Midjourney model is best for generating psychedelic art.
- 🌟 **High-Resolution Images**: The generated replications are described as high-resolution and complex, with a focus on architecture.
- 🏢 **'Cute' and 'Structure' Keywords**: The words 'cute' and 'structure' are emphasized in the prompt to generate non-evil and non-anthropomorphic entities.
- 🔄 **Multiple Rerolls**: It often takes several rerolls to get a satisfactory result.
- 🔍 **Zoom for Visual Glitches**: When upscaling, look for visual glitches like opacity overlays and smudges.
- 🌈 **Colorful Prompts**: The speaker avoids using 'psychedelic' in prompts, instead opting for 'colorful' to prevent unwanted visual effects.
- 🎭 **Avoiding Psychedelic Art Clichés**: The goal is not to create psychedelic art but to replicate the psychedelic experience using AI.
- 🔧 **Adjusting Magnific Settings**: Settings like creativity, HDR/detail, and resemblance are adjusted to refine the generated images.
- ➗ **Practicality Setting**: Setting the practicality to zero can help in reducing visual errors and enhancing the final output.
- 🔄 **Iterative Process**: The process involves running the image through the upscaling multiple times to improve clarity and coherence.
- 🚫 **Not for Art or NFTs**: The speaker clarifies that the generated replications are not intended for commercial art or NFT sales.
Q & A
What is the main topic of the tutorial?
-The main topic of the tutorial is creating high-resolution, complex DMT hyperspace AI-generated replications.
Why is Mid Journey model version 5 preferred for generating psychedelic art?
-Mid Journey model version 5 is preferred because it better captures the desired vibes for psychedelic art, even though newer versions are technically more sophisticated.
What is the significance of the word 'cute' in the prompt?
-The word 'cute' is used to offset the potential for generating entities that look unnecessarily evil and scary, resulting in a more nonhuman or non-anthropomorphic appearance.
What is the role of the word 'structure' in the prompt?
-The word 'structure' is used to focus the generation on architectural aspects, which is the current interest of the tutorial presenter.
How does the presenter ensure the generated entities are not scary?
-The presenter includes the word 'cute' in the prompt to ensure the generated entities have a more friendly appearance.
What is the process for generating the replications?
-The process involves setting the Mid Journey model to version 5, using a specific prompt with focus words like 'cute' and 'structure', and then upscaling the generated image using Magnific with a detailed prompt and specific settings.
Why is the creativity level set to five and the HDR or detail increase to eight in Magnific?
-The creativity level and HDR or detail increase are set to these values to achieve a complex and intricate interdimensional hyperspace structure in the generated image.
How does the presenter avoid visual errors during the upscaling process?
-The presenter adjusts settings such as reducing the practicality level and modifying the prompt to be more specific about the desired outcome, and re-running the image through the upscaling process with reduced settings if visual errors are detected.
Why does the presenter avoid using the word 'psychedelic' in the Magnific prompt?
-The presenter avoids using 'psychedelic' because it tends to result in a style that looks more like tie-dye or Paisley patterns, which is not the desired look for the replications.
What is the presenter's goal with these AI-generated replications?
-The presenter's goal is to replicate the psychedelic experience using artificial intelligence, viewing it as a phenomenological or scientific endeavor rather than an artistic one.
How does the presenter refine the generated image to remove incoherency?
-The presenter reuses the final image, reduces the creativity and HDR settings, and runs the image through the upscaling process again to refine the details and clarity.
What is the final outcome of the tutorial?
-The final outcome is a high-resolution, complex DMT hyperspace replication that captures the desired psychedelic experience, which the presenter then shares on their page.
Outlines
🎨 Generating High-Resolution Psychedelic Art with AI
The speaker discusses their recent tweet showcasing high-resolution, complex DMT hyperspace AI-generated replications. They claim to have discovered an easy technique for creating these psychedelic art pieces. The speaker shares their enthusiasm for the quality of the generated images, particularly one featuring architecture. They explain that while the example focuses on architecture, the technique can be used to generate various aspects of high-level psychedelic states, including entities, geometry, and different rooms. The speaker emphasizes the importance of using the word 'cute' in their prompt to avoid generating entities that look evil or scary. They also mention using the Mid Journey model version 5 for the best results and provide a detailed description of their prompt, which includes the words 'cute' and 'structure'. The speaker demonstrates the process of generating and upscaling the images using Discord's Mid Journey and an AI tool called Magnific, adjusting various settings to achieve the desired outcome.
🔍 Refining AI-Generated Hyperspace Images for Visual Coherence
The speaker continues by focusing on the process of refining the AI-generated images to achieve visual coherence and high resolution. They mention looking for visual glitches during the upscaling process, such as opacity overlays and smudges, and adjusting the 'practicality' setting to minimize these issues. The speaker also discusses the importance of using colorful prompts to generate vibrant hyperspace structures, avoiding the term 'psychedelic' as it can lead to undesirable artistic styles. They demonstrate how to address visual errors by reusing the final image, reducing the creativity and HDR settings, and running the image through the upscaling process again. The speaker concludes by emphasizing their goal of replicating the psychedelic experience using AI, rather than creating traditional art or NFTs, and invites viewers to explore their page featuring various entities and architectural designs generated through this process.
Mindmap
Keywords
💡DMT
💡Midjourney
💡Hyperspace
💡Entities
💡Architecture
💡Prompt
💡Magnific
💡Upscaling
💡Psychedelic States
💡Creativity
💡Practicality
Highlights
The speaker has developed a consistent and easy technique for generating high-resolution, complex DMT hyperspace AI replications.
The generated replications are considered some of the best the speaker has personally seen.
The technique can be used to generate various aspects of high-level psychedelic states, including entities, architecture, geometry, and different rooms.
The Mid Journey model version 5 is recommended for generating psychedelic art.
The prompt used for generating the replications includes the keywords 'cute' and 'structure', with 'cute' used to offset potentially scary entities.
The speaker emphasizes the importance of the prompt's wording, as it significantly influences the outcome of the generated images.
The process involves several rerolls to achieve the desired result.
Upscaling the generated image using a tool like Magnific is a key step in the process.
The scale factor is typically set to two, and the prompt on Magnific is less critical than the one used in Mid Journey.
The speaker suggests using a prompt focused on 'complex, intricate, interdimensional hyperspace structure' for generating intricate patterns.
Adjusting settings such as creativity, HDR/detail, and resemblance can fine-tune the output.
Practicality is often reduced to minimize visual glitches in the upscaled images.
The speaker avoids using the term 'psychedelic' in the Magnific prompt to prevent an undesirable artistic style.
Visual coherence is checked for errors that may indicate AI mishandling rather than intentional artistic choices.
If visual errors are present, the image is reprocessed with reduced creativity and HDR settings to improve coherence.
The final output aims to replicate the psychedelic experience using artificial intelligence, rather than creating traditional art.
The speaker is not interested in selling NFTs but is more focused on the phenomenological and scientific aspects of the endeavor.
The generated replications are showcased on the speaker's page, featuring a variety of entities and architectures.