From Midjourney to Cinema4D v.1

In Ai.Grafik I explore the relationship between numerous typologies and styles of Art and Design (generative art, code art, abstract art, particle and fluid systems, graffiti, light, architecture, sketches, industrial, cubism, neoplasticism, art déco, minimalism, sci-fi, steampunk, dieselpunk, soft volumes) with typography using Gen-AI tools.

As I create new compositions, I select some I can use as a basis to get the desired result and design enhanced 3D versions and animations. This way, I test how we can get finer outcomes using Gen-AI tools as an inspiration and creativity booster tool inside our workflow.

I present the first 3D model: an A-letter sculpture in neoplasticism style.

Regarding the methodology, I used MJ (5.2) in the initial stages of the project, searching for a proper prompt structure (definition, shape, style, camera, illumination, materials, etc.), intending to get promising (but not final) results in neoplasticism-style typographic compositions. Afterwards, I refined the outcome until I got the desired one, or at least one closest to the expected one.

In the (short) future, MJ will present functional improvements that permit better outcomes in images/animations and tangible variations that are now impossible to achieve. Currently, it can make astounding variations of one composition. Still, it generally fails when it needs to set or change specific/exact variables of a (for example) typographic 3D render: a specific camera angle or DoF, particular material properties, or modify ambient occlusion.

This fact leads to one of the risks of this type of image generation: being satisfied with a visually aesthetic result but giving up on what the project needs. The designer yields his intention to what the IA can generate.

Then, it is where MJ must give way to the functional design.

Following the AI image, I created the 3D model in Cinema4D/Redshift, making an almost identical copy and improving aspects to get the desired quality, such as materials, camera DoF, realistic details, imperfections, geometry, etc.

Once satisfied with the model, I made a short opening animation.

https://www.youtube.com/watch?v=jJaIWxl_7SQ

To sum up, I am amazed to verify how MJ helps to get the expected results. A well-trained MJ designer can generate valuable content regarding inspiration and experimentation in the early stages of the project. The downside is the risk of letting MJ decide on an eye-candy outcome that fits neither the designers' will nor the project requirements.

Incorporating and using these tools offers an advantage that is still difficult to appreciate due to the speed of their evolution. I consider it a must.