Integrating AI into my 3D workflow
Oh look! Someone using Midjourney! Gather around as I force an AI to produce image after an image using text prompts… WAIT! Don’t leave, I promise this might be interesting!
In an effort to ride the tsunami of AI art instead of fall under it, I wanted to see how a tool like this would fit into my workflow. That being style frame and concept ideation.
While drawing with my 4 year old we came up with this little guy. A cactus holding a balloon is adorable and fraught with danger for the balloon! (I did the line art he supplied the colour).
Now time for the magic AI to create exactly what’s in my hea… Ahh… not quite right! This was my prompt for those playing at home. “a cute cactus done in a kawaii style 3D render, on a wooden table” Now to try out the new engine released…
Awwww! OK now we’re talking!! Exactly the same prompt, but using the newly released V4 engine, the engine is improving exponentially and is such a powerful tool!
Development
Now I had some style frames going (there are many, many more). I dialled in on a few things I knew I wanted in the end product: Saturated colours, shallow depth of field and a claydough / stop motion style to pay homage to the origins of the idea coming from my 4yo.
I started with some simple box modelling, and getting the general look fleshed out. I know I wanted to have him rigged up so being patient at this stage can save some headaches down the road.
Now it’s on to texturing, thankfully Greyscale Gorilla released a clay texture pack that didn’t take too much work to get him to look like playdough.
Here we have the textures, lighting and composition dialled in. I wanted to add some known office supplies around our Cactus to sell the scale of him, and using a shallow depth of field this is reinforced.
In the end I went with bubbles instead of a balloon as one balloon popping is sad, but 50 bubbles popping is playful.
I hope you enjoyed this peek into what goes on behind the scenes making a cactus laugh at bubbles popping.