Mac Baconai
patreon
WINDS OF THE NEW ERA: WORKFLOW & CRITIC
🕑 Added 2024-08-15 23:53:51 +0000 UTCHey Team, grab your coffee, and put on some good music. This is going to be a long and information-packed article. Some of you are quite advanced in generative AI, while others are just learning. Don't worry; you'll find some crucial insights in this article, especially about Midjourney, no matter your level. The article will be divided into two parts:
Music Link: Good start for the track list. Also this piece has great video for our article context.
''when you at it from up here you get an appreciation of our world is a beutiful place and we do need to take care of it.''
1-Workflow: First, I will explain my general design approach and how I created the posts I've recently shared, detailing it through an example prompt and workflow. In this section, we'll explore how to approach tools like Midjourney, Leonardo, Photoshop, and others. Combining different programs can have powerful advantages. Each tool has its strengths and weaknesses, and new ones are constantly emerging. Therefore, it is crucial to stay updated and follow developments. For me, the tools that currently create the most powerful synergy are:
Midjourney + Leonardo + Photoshop + Lightroom.
2-Wind of Change: In the second section, we'll take a brief look at the evolution of art through examples from the past and present. This part will be a bit like an art history lesson, and I'll try to keep it as short as possible. If you find it interesting, I strongly recommend doing further readings. A strong prompt input should also draw significantly from the intellectual heritage and art movements of the past. As much as the production process, your ability to critique your outputs and select the most effective ones is essential. The knowledge you gain at this stage will be quite important for your -p code. (what is --p code)
-----------------------------------------------------------------------------------------------------------------
!!!BEFORE WE START: If you don't have enough experience with Midjourney and want to achieve quick and effective results, I recommend paying close attention here. If you have experience with Midjourney you can skip this part.
I've been actively using Midjourney from the very beginning. The biggest leap in model infrastructure occurred with v5. Therefore, working with models from v5 to v6.1 (the most current model at the moment) will be sufficient. You can use Midjourney in two different ways: via the Website or Discord. Each has its own advantages, and I actively use both. On the website, you can generate images, access your prompts, and create separate collection-folders for your works. For Discord (optional), you can set up your own channel and connect the Midjourney Bot there. This allows you access to some additional commands (/describe, /blend, /tune, etc.). In both scenarios, Official Midjourney Documentation Website should be your essential handbook when starting out.
''Our archive is our temple,
Our prompts are our sword,
Our parameters are our shield,
In the canvas of creation, we wield.
Midjourney guides our artful hand,
Where visions flourish, and dreams expand''
In short we need to focus on these 3 title and their combinations to get fascinating results:
1-Archive
2-Prompts
3-Parameters
If you want to practice from the zero; you can my check my two education bundles, these are designed carefully with unique topics including their prompts in alt. These have v5 prompts so you can get better results with v6 and v6.1. In any case these have quite importing prompting philosphy/experiments for my design approach. I strongly recommend you to explore.
Link: MIDJOURNEY BUNDLE VOL.1 (Prompts in ALT)
Topics
1-Traditional Patterns and Cultures
2-Architectural Artifacts and Gemstones
3-Banner Design and Showcase
4-Pop A'rt and Blending
5-Prompt Alchemy and Creation Spell
6-Two Emoji Photorealism and Storytelling
Link: MIDJOURNEY BUNDLE VOL.2 (Prompts in ALT)
Topics
1-Zoom Out Feature and Blending
2-Academai & Content Strategy
3-Cliffside Modernism
4-Reconquista Major Arcana Tarot
5-Anchor Method & Promtum Physics
6-Surrealist Blending & Inpainting
SUBSCRIPTION PLANS: MIDJOURNEY & LEONARDO
A-Midjourney has four plans: Basic($10), Standart($30), Pro($60), Mega($120)
*These are monthly prices %20 off with annual
If have no experience at all, Basic Plan should be fine for you. But I strongly recommend the Standart Plan for a fresh start. (Basic Plan has Limited generations (~200 / month), and it's not enough imo. If you choose Standart Plan you will get Unlimited Relaxed generations.)
I'm using Pro Plan. It's enough for me. You can start with Standart then you can upgrade when you feel ready. Upgrading will provide more fast jobs and stealth image generations.
B-Leonardo is also great tool with lots of feature. I'm mostly use it for Universal Upscaler to increase the quality of Midjourney outputs. You can try and check their plans in their website.
On the other hand. Leonardo gives you 150 Free credits daily. One image to upscale costs 30 credits. So It's quite enough to upscale 5 image per day.
Also here is my MJ image generation data;
Lifetime Usage: 78133 images
Fast Usage: 28904 images
Turbo Usage: 28 images
Relaxed Usage: 49201 images
P.S. The models and prompt consistency of Midjourney vary from model to model. It is possible to achieve impressive results through different prompt techniques and linguistic experiments. Alternatively, models like Stable Diffusion can certainly be used. However, their setup is more solid, and the learning curve process is entirely different. While models like Stable Diffusion offer you more control, Midjourney provides a more powerful experimental environment for achieving strong aesthetics. Once you build a strong archive full of aesthetic curation. It will be easier to manage the process and produce unique outputs.
-----------------------------------------------------------------------------------------------------------------
1-WORKFLOW
Now we can review the workflow of my recent works with example visuals. Before we start, I'd like to define a term: Core/Creator Prompt
Core/Creator Prompt: Powerful prompt designed for a specific purpose with defined compartments. It should produce consistent results within the designated theme. We should be able to use it quickly and effectively with minor adjustments.
In most of my recent works, I used the following Core/Prompt. This Core/Creator Prompt has proven effective due to its ability to maintain consistency and adapt to different themes with minor adjustments. (Occult Forest Fashion is also designed as a Core/Creator prompt, but it has a completely different prompt structure.*)
Core Prompt:
wide angle shot enormous landscape of fantasyland vistas and magicians riding their horses to source of energy beam of the divine light, enigmatic flame of crimson anger fire explode, ancient gigantic creature dividing the earth, with background fascinating nature, cinematic shot of wide view editorial fiction film of the hanging gardens of ruins of ancient kingdom, a multi-layered civilization with many people is standing in the middle of the great rocky-tuff valley of high council that is architecture inspired ancient city, chamber of the free people are spoken, radiate happiness with the light of wisdom, columns from the river sustain organic rocks. Ionic plants born from the soil merge with the natural rock walls of the living city, unreal engine, cgi, octane render, highly detailed,in the style of hyper-realistic sci-fi, matte painting, gray and crimson, zbrush, monochromatic chaos, photo-realistic techniques, award-winning --chaos 33 --ar 5:6 --style raw --stylize 1000 --v 6.1 --sref 3271874953
The prompt structure consists of 4 main sections:
(You can check the second image I uploaded for the highlighted core prompt.)
Green Part: This is the most effective and fastest-working part. I generally made adjustments to this part when changing scenes.
Yellow Part: This part is where details are added. The purpose here is to note the details you want to see. You can weight the tokens you need as ::X (X=number). This area helps you select the tokens you want to highlight in the composition more easily.
Blue Part: I did not change this part at all.
Red Part: Definitely the most important part. This area is where you have the most control over the overall aesthetic. Please check the comparison images.
You can check and compare the result with images. Also here the most important parameters that I use:
--chaos : 33 is optimum value for my general workflow. This provides you more different composition with prompted theme.
--ar : defines the image ratio
--style raw : adds photorealism
--stylize : increases details
--v 6.1 : selected model version
--sref : unique code for specified aesthetic (here this sref provides colors and floral patterns, It's quite powerful so you should try it with your own prompts.)
Also please don't forget to read Official Midjourney Documentation Website to learn each code's full potential. You can easily search the parameters there.
Finally, I slightly added more parameters at images to get comparison results so you can easily check them!
WHAT IS --P CODE & WHY IS THE --P SO IMPORTANT?!
--personalization is the new feature of Midjourney. You should go to midjourney and rank the images you like. (More info: Link) --p is optional but it has a significant advantages to express your unique style.
This will train you aesthetic. So your selected images is extremely important.
That's Why:
1.DO NOT RUSH, SKIP IF YOU DON'T LIKE THE IMAGE
2.FOCUS ON HIGH VALUE/AESTHETIC/ART MOVEMENTS
3.TRY TO FIND UNIQUE ONE
4.DO NOT SHARE YOUR --P CODE (Because, you'll always work/train with it later)
5.TRY YOUR P
6.STOP WHEN YOU FEEL OKAY
It's so personal actually. Decision you made will affect your later aesthetic with --p. You can always open/close it.
How to find the unique --sref code!?
I have a quick solution for this.;
Use this prompt:
show me the power of this scene --c 33 --ar 5:6 --sref random --s 1000 --v 6.1
1.Run the prompt
2.Reuse the prompt when you feel to find something interesting/powerful
3.Use the selected code 3 time and run the prompt with same code
4.Check the overall aesthetic, color correction, atmosphere
5.If this collection tells a strong story to you. You are ready to use this code.
P.S. do not use --style raw at this step, while exploring new --sref codes. we don't need realism. we are tracking the emotion.
Also, A fun website to get quick feedback and data: AI Art Critic
You can upload your image to get quick feedback.
LEONARDO UNIVERSAL UPSCALER
It's quite simple process.
1-Go to Universal Upscaler
2-Upload an image
3-Select a model (I mainly use Cinematic 5-7)
You can check my post to learn more: Link
2-WIND OF CHANGE
You can skip this part if you're not interested. It mainly concerns architecture, art, and AI critiques. However, having intellectual knowledge about art movements and your professions can provide significant advantages. Instead of focusing on easy prompts, try to learn more and work with different themes.
I joined the Charis competition and participated in all categories with a total of 99 images. You can check out my portfolio below. Experimenting with various categories, especially architecture, visual arts, fashion, and photography, will help you build a strong archive and enhance your intellectual background
Portfolio: https://charis.ai/macbaconai
2.1. ARCHITECTURE & MOVEMENTS
I will give you some examples how the changing situations affect the archi-movements in history.
World War I and Art Deco: The Birth of Modernism
After World War I, societies were deeply affected and sought renewal and modernity. This quest led to the rise of Art Deco, a design movement that embraced new industrial possibilities while rejecting the dullness of industrialization. Art Deco, with its geometric forms and elegant style, became a symbol of the 1920s and 1930s sophistication. The new materials and techniques from the Industrial Revolution were crucial in creating Art Deco’s distinctive buildings, furniture, and art.
Art Deco not only filled the gap left by the war but also took advantage of new technologies, driving significant progress in art and architecture. (Image Source: Link)
The Bombing of Berlin and Modernism’s Rise
After World War II, Berlin was heavily damaged and needed to be rebuilt quickly, which gave rise to Modernism. Unlike Art Deco’s decorative style, Modernism focused on simplicity, functionality, and practicality. With a growing need for housing, Berlin’s reconstruction used prefabricated and modular techniques, leading to the creation of a new city that balanced design with practicality. Modernism became a symbol of recovery, offering practical solutions that helped society rebuild.
(Image Source: Link)
The Emergence of Postmodernism
By the 1970s, Postmodernism emerged as a reaction to Modernism's minimalism. It brought back historical styles and ornamentation, adding diversity and historical references to architecture, as seen in cities like Berlin. Postmodernism, with its eclectic and innovative designs, redefined architectural identity by combining Modernism’s functionality with a renewed appreciation for history and aesthetics. (Image Source: Link)
AI and New Artistic Movements
Just as the Industrial Revolution and post-war optimism led to Art Deco and Modernism, today’s technological advancements and AI tools are paving the way for new artistic movements. In a world facing growing risks, the creative and hopeful use of AI could be key in shaping the future of art and architecture. AI-driven design tools, much like past technological innovations, are setting the stage for a new aesthetic in art and architecture, with great potential to build the cities and communities of tomorrow. The optimism of Art Deco and the practicality of Modernism, combined with AI’s possibilities, could usher in a new era of creativity and transformation.
2.2. AI ART CRITIC
Here, I will write a strong art critique and manifesto against those who reject AI, with examples;
*you can also check the visuals from the slide.
The True Evolution of Art and the Rising Shadow of AI: Those Who Resist Innovation Are Doomed to Fail
We are tired of hearing narrow-minded and reactionary arguments about what art is. Claiming that AI-generated content is not art is merely a sign of cowardice and resistance to innovation. If Andy Warhol’s reworking of existing images is considered art and Marcel Duchamp’s presentation of a urinal as a piece of art was revolutionary, then AI-generated works are also art!
Art has always been an expression that demands innovation and pushing boundaries. When Marcel Duchamp presented a ordinary urinal as “Fountain” to the art world in 1917, he reshaped the definition of art. Duchamp showed that art is not just a physical object but an expression that gains meaning through the artist’s vision and context. His bold move freed art from narrow confines and provided a broader perspective.
Andy Warhol, by introducing pop culture into the art world, presented everyday objects and cultural symbols as part of art. With works like Campbell's Soup Cans, he demonstrated that art could be a reflection of mass culture and that ordinary objects could carry profound artistic value. Warhol led the democratization of art and its accessibility by merging mass production techniques with art.
Today, claiming that AI-generated content should not be considered art reflects a misunderstanding of Duchamp and Warhol’s conceptions of art. Art continues to evolve with innovative tools and ideas. AI supports this evolution by offering new creative avenues for artists. Accepting that art is not confined to narrow definitions is about understanding past achievements and innovations.
However, it is also important to maintain the value of human creativity and originality in the face of AI’s rising influence in the art world. While AI-generated works may mimic the essence of art, they lack the internal creativity and personal emotional depth. Duchamp and Warhol’s works reflected the artist’s personal vision and aesthetic decisions; this remains something AI’s limited creative processes cannot fully achieve. Yet, this gap can be addressed through strong curation and complete mastery of AI by the artist.
Art is not only about aesthetic values but also a reflection of human experience, thought, and emotion. The tools AI offers cannot replace these depths, and it is crucial to emphasize the importance of human creativity to understand art’s true potential. While those who resist innovation will be buried in the dusty pages of history, the brave will shape the future.
Those who refuse to recognize AI-generated works as art are those who do not understand art. They are doomed to fail. We will continue to create, innovate, and redefine the boundaries of art. Let’s celebrate the evolution of art and focus on the future by embracing new tools, ideas, and perspectives. AI’s innovative capabilities will play a crucial role in the next wave of art, expanding the boundaries of the new era.
CONCLUSION
In summary, exploring generative AI tools like Midjourney and Leonardo offers exciting possibilities for artists. By understanding the technical aspects, experimenting with different prompts, and staying updated on advancements, you can enhance your creative practice and produce unique and impactful works.
Please feel free to ask if you need something! I hope you find this article helpful.
Comments
Pedro Audirac
"Art has always been an expression that demands innovation and pushing boundaries [...]" there is nothing more to say.
academai
Hey Felipe, what kind of content do you want to see? Actually I'm waiting V7 for MJ. But feel free to drop if you have something in your mind!
Felipe
Do you plan on bringing new content here?
