Harnessing the Power of AI: Crafting Breathtaking Images with Midjourneyand AI

mansour ansari
May 28, 2023
5 min read

Harnessing the Power of AI: Crafting Breathtaking Images with Midjourney AI and How AI Strolled into My Life and Turned It into a Technicolor Dream!

You know that moment when your life is like a black-and-white movie, and suddenly, Technicolor kicks in? That, my dear friends, is precisely what it felt like when AI traipsed into my life. And not just any AI, mind you, but the astonishing world of image and illustration generation!

From the frosty month of January 2023, I started moonlighting as an AI detective, burying myself in the mysteries of this fascinating world. Why, you ask? Two reasons: First, I had my socks knocked clean off by AI's image and illustration generation. And second, its buddy-cop partnership with quantum computing left me wide-eyed and slack-jawed!

So today, I've decided to spill the beans. I'm opening my diary of AI adventures and sharing some of the nuggets I've dug up along the way. I'll also be sharing some nifty tricks on how to whisper sweet nothings into the AI's ear, so it produces the exact images that you dreamt of last night or thought up while indulging in your character-building escapades.

So buckle up, folks, as we embark on this rollercoaster ride of my AI experience. You're about to learn how an ordinary Joe like me managed to tame the extraordinary beast that is AI!

In the world of creativity and imagination, there are few boundaries left untouched, and AI is proving to be a valuable tool in the realm of image generation. I recently embarked on a creative journey with Midjourney AI, and the results have been nothing short of stunning.

Here's an example: the creation of a character named "Zagros." Zagros was an idea, a figment of imagination, a character I am developing for a sci-fi comic book I like to produce. But with a carefully crafted prompt, Midjourney AI brought him to life in high-resolution, cinematic images.

I now have 2 different candidates for the character I am building. As you see, each image has a different facial expression.

Which one do you think is THE ONE? See the top image.

So, the prompt I used was as follows:

"Epic Cinematic action scene of a hero, detailed UHD, surreal vibes, by atelier olschinsky 8k, photorealistic, hyper-realistic, cinematic photo, epic composition, color grading, unreal engine, hdr, octane render, depth of field, hyper-detailed, intricate details, lens flare, Shot on hasselblad h6d-400c, f/22, Shutter Speed 1/1000, lens LEICA 90mm, RAW, megapixels, cinematic scene, epic composition, super resolution, tone mapping, ray tracing, FKAA, TXAA, RTX, SSAO, cgi, vfx, sfx, Megapixels, --q 2 --v 5.1 --ar 16:9 --s 750 --stylize 300." This might seem like a lot to take in, but let's break it down. The prompt begins by setting the scene: "Epic Cinematic action scene of a hero." It goes on to specify the level of detail and style: "Detailed UHD, surreal vibes, by atelier olschinsky 8k, photorealistic, hyper-realistic." It sets the tone and mood with words like "epic composition, color grading, unreal engine."

The prompt then ventures into camera specifics, simulating a real-life photoshoot. It outlines the camera type (Hasselblad h6d-400c), aperture (f/22), shutter speed (1/1000), and lens type (LEICA 90mm). It calls for a RAW format, signaling the best quality and most flexible format for post-processing.

Then we get into the realm of digital enhancements: "Super resolution, tone mapping, ray tracing, FKAA, TXAA, RTX, SSAO, cgi, vfx, sfx, Megapixels." These terms refer to various digital technologies and techniques used to enhance image quality, providing hyper-realistic details and effects.

Finally, it ends with a series of commands that dictate the quality (--q 2), version (--v 5.1), aspect ratio (--ar 16:9), style (--s 750 --stylize 300) of the final image.

The result? A breathtaking image of a hero named Zagros, exactly as imagined. This is the power of AI image generation. With the right prompts, the possibilities are infinite. Just remember, crafting the perfect prompt is an art in itself. It's a collaborative effort where your imagination meets the AI's processing power. So, go on and create, experiment, and let the magic of AI amaze you!

So, while playing around with AI, I also caught these images generated with the AI with a similar prompt using a similar technique. I am also including the complete meaning of these parameters.

The breakdown of this prompt:

Atelier Olschinsky: This refers to a professional photography studio known for its high-quality, visually stunning images.

8k, UHD, Photorealistic, Hyper-realistic: These terms indicate the high resolution and the level of detail expected in the generated image. 8k and UHD (Ultra High Definition) are terms that describe the resolution of the image. Photorealistic and hyper-realistic suggest that the image should closely resemble a high-quality photograph.

Cinematic Photo, Epic Composition, Color Grading: These terms describe the style and mood of the image. "Cinematic" suggests a dramatic, film-like quality. "Epic composition" and "color grading" point to a grand, carefully composed image with specific color tones.

Unreal Engine, HDR, Octane Render: These are tools and techniques used in computer graphics. Unreal Engine is a game engine used for creating 3D graphics, HDR (High Dynamic Range) is a technique used to enhance the contrast and color of an image, and Octane Render is a popular rendering engine.

Depth of Field, Hyper-detailed, Intricate Details, Lens Flare: These terms refer to specific visual effects. Depth of field relates to the focus of the image, "hyper-detailed" and "intricate details" suggest an extremely high level of detail, and "lens flare" is a specific lighting effect.

Shot on Hasselblad H6D-400c, f/22, Shutter Speed 1/1000, Lens LEICA 90mm, RAW: These specifications mimic a professional photo shoot setup costing big bucks! . Hasselblad H6D-400c is a high-end camera, f/22 is the aperture setting, 1/1000 is the shutter speed, LEICA 90mm is the lens used, and RAW indicates the image format, which contains minimally processed data from the image sensor.

FKAA, TXAA, RTX, SSAO: These are advanced graphics rendering techniques. FKAA stands for Fast Approximate Anti-Aliasing, TXAA is Temporal Anti-Aliasing, RTX refers to Nvidia's ray tracing technology, and SSAO stands for Screen Space Ambient Occlusion.

Cgi, vfx, sfx: These abbreviations stand for computer-generated imagery, visual effects, and sound effects, respectively. In this context, they probably refer to the desired quality level of the image.

--q 2 --v 5.1 --ar 16:9 --s 750 --stylize 300: These are the command-line arguments used to fine-tune the image-generation process. They control the quality (--q), version (--v), aspect ratio (--ar), size (--s), and style (--stylize) of the generated image, respectively. I am still learning!

As you can see, this is quite an in-depth prompt designed to extract as much detail and quality from the AI as possible! I can't wait to learn more about how to make my AI friend bring out the inner artist in me!