Military Robot Sculpt to Ai
Another one that started as a hard surface sculpt to test out shapes (Image 1). I wanted this one to have an almost spider like quality in terms of the head shape, the eyes having full wrap around for positioning on the battlefield. I created a few Ai variations, some really interesting results with the mandible pods around the chin, looking more like a spider.
However I really liked the original sculpt and wanted to keep close to the design and just polish areas that were only really 'block out shapes'.
Image 1
3D-Coat isn't the best render out there, but when combining it with the pbr shaders and a few positioned lights it gets the job done quickly. As the final output is not the render, its not too concerning (however would be great to able to to use a better render engine for final output, something where I didn't need to leave the 3DC environment).
Image 2
As you can see (Image 2) the final Ai output didn't deviate too much from the original sculpt. I did post work in Photoshop, cherry picking variations. My prompt also included 'battle damaged' as I wanted to add those details to the mix, rather than having a clean finish. Its a great way to decide on those aspects quickly rather than sculpting.
The idea here is I have a base sculpt with all the main landmarks ready to go. I then drop them into Ai to test how I want to refine those base features. I make any adjustments needed in PS and then I'm ready to transfer these back onto the base sculpt.
'Why not just sculpt the whole thing, retopologise and the bake those 'battle damaged' areas in Substance or 3DC'? - Well you could certainly do that, but I see that as a total commitment to the design and why do that? What if you could find that golden nugget? The block out doesn't take much time at all and I'd rather use that time gained to explore the options using Ai. I feel this allows you to jump into 3D sketching (sculpting) first and then refines using 2D Ai to prep the design fully for committing.
Once that's done, I have a prompt, a seed and also a sculpt which I can re-position, re-light, use different materials and in a matter of seconds have another angle to revise (Image 3)
Image 3
Is the design exactly the same - no. However in a short amount of time, I'm taking a design and rapidly building a picture in a lot of detail as to the direction of this character. For example if I prefer some of the front view design elements - I could build these iterations into the base sculpt, which solidifies these details (almost like baking).
The future of this process will be the ability to move between 3D and Ai seamlessly where the Ai can use a 3D base shape and mold it based on the designers prompt. The designer remains in control with traditional sculpt tools and Ai sculpting prompts.
However I really liked the original sculpt and wanted to keep close to the design and just polish areas that were only really 'block out shapes'.
Image 1
3D-Coat isn't the best render out there, but when combining it with the pbr shaders and a few positioned lights it gets the job done quickly. As the final output is not the render, its not too concerning (however would be great to able to to use a better render engine for final output, something where I didn't need to leave the 3DC environment).
Image 2
As you can see (Image 2) the final Ai output didn't deviate too much from the original sculpt. I did post work in Photoshop, cherry picking variations. My prompt also included 'battle damaged' as I wanted to add those details to the mix, rather than having a clean finish. Its a great way to decide on those aspects quickly rather than sculpting.
The idea here is I have a base sculpt with all the main landmarks ready to go. I then drop them into Ai to test how I want to refine those base features. I make any adjustments needed in PS and then I'm ready to transfer these back onto the base sculpt.
'Why not just sculpt the whole thing, retopologise and the bake those 'battle damaged' areas in Substance or 3DC'? - Well you could certainly do that, but I see that as a total commitment to the design and why do that? What if you could find that golden nugget? The block out doesn't take much time at all and I'd rather use that time gained to explore the options using Ai. I feel this allows you to jump into 3D sketching (sculpting) first and then refines using 2D Ai to prep the design fully for committing.
Once that's done, I have a prompt, a seed and also a sculpt which I can re-position, re-light, use different materials and in a matter of seconds have another angle to revise (Image 3)
Image 3
Is the design exactly the same - no. However in a short amount of time, I'm taking a design and rapidly building a picture in a lot of detail as to the direction of this character. For example if I prefer some of the front view design elements - I could build these iterations into the base sculpt, which solidifies these details (almost like baking).
The future of this process will be the ability to move between 3D and Ai seamlessly where the Ai can use a 3D base shape and mold it based on the designers prompt. The designer remains in control with traditional sculpt tools and Ai sculpting prompts.