Thread
Big-budget games & movies from God of War to Avatar are brought to life by armies of 3D artists
Today: these artists are technical wizards yielding complex tools🧙‍♀️
Future: creativity will be unleashed at the speed of thought👩‍🎨
How Generative AI will transform 3D artist job's👇
Today: these artists are technical wizards yielding complex tools🧙‍♀️
Future: creativity will be unleashed at the speed of thought👩‍🎨
How Generative AI will transform 3D artist job's👇
Games sit at the bleeding edge of technology x art. To design and produce a game of beauty requires a unique blend of left brain (logic; reasoning) and right brain (creativity; imaginative) juices
The same holds for linear, pre-rendered 3D creation (animation, VFX) where artists use many of the same creation tools
In fact, the Pixar campus is laid out like a brain: engineers, graphics researchers on one side; storytellers, designers and non-technical artists on the other
In fact, the Pixar campus is laid out like a brain: engineers, graphics researchers on one side; storytellers, designers and non-technical artists on the other
Within a large studio, artists are highly specialized and wield unique tools to accomplish their tasks.
Distinct roles include:
- Art Director
- Concept artist
- Gameplay designer
- Narrative
- Environment artist
- Character artist
- Animator
- Motion capture artist
Distinct roles include:
- Art Director
- Concept artist
- Gameplay designer
- Narrative
- Environment artist
- Character artist
- Animator
- Motion capture artist
- Scanning artist / technician
- Rigger
- Prop artist
- Texture artist
- Lighting artist
- V/FX artist
- UI designer
- Technical artist
- Audio designer
- Level designer
- Layout artist
.
.
.
- Rigger
- Prop artist
- Texture artist
- Lighting artist
- V/FX artist
- UI designer
- Technical artist
- Audio designer
- Level designer
- Layout artist
.
.
.
This specialization results from the need to learn complex tools and tricks to master the job at hand
Environment Artist -> @SpeedTreeInc
Animation -> MotionBuilder, @Mixamo
Modeler / Character Artist -> Zbrush
Texture Artist -> Substance
Audio Designer -> FMOD, WWise
...
Environment Artist -> @SpeedTreeInc
Animation -> MotionBuilder, @Mixamo
Modeler / Character Artist -> Zbrush
Texture Artist -> Substance
Audio Designer -> FMOD, WWise
...
But also the rigid, assembly line like nature of production pipelines. Here's a typical VFX pipeline:
Game production pipelines have always been more iterative, given the ability to compile and render locally in real-time but yet the specialization remains
Game production pipelines have always been more iterative, given the ability to compile and render locally in real-time but yet the specialization remains
With Generative AI, much of the technical complexity for an artist will be abstracted away so that they can produce better content, faster and cheaper
What does that mean?
What does that mean?
Artistry will become more about expressing creative intent rather than laboriously painting pixels in sophisticated tools. Artists will be generalist creative directors rather than specialized technicians.
How?
How?
Empty scene with gray cube be-gone!
Text-to-3D model generation will give every artist an initial starting point to work from, only limited by their own imagination. No more empty canvases. No more waiting on the one team modeler to finish her assignment
Text-to-3D model generation will give every artist an initial starting point to work from, only limited by their own imagination. No more empty canvases. No more waiting on the one team modeler to finish her assignment
Assisted artistry features to edit to perfection
Time intensive, non-creative tasks will be automated away. Rather than grinding through tedious tasks like UV unwrapping, retopology and texture painting, right click and describe what you want directly
Time intensive, non-creative tasks will be automated away. Rather than grinding through tedious tasks like UV unwrapping, retopology and texture painting, right click and describe what you want directly
Creation "Auto-Complete"
Once you’re happy with your "hero" asset(s) - the vocal point of the scene - you'll “autofill” all the other assets in the scene around it based on semantic understanding of the existing assets. Then refine as needed
Once you’re happy with your "hero" asset(s) - the vocal point of the scene - you'll “autofill” all the other assets in the scene around it based on semantic understanding of the existing assets. Then refine as needed
Obviously, today's early PoCs have a long way to go before they become truly transformative for professional artist production workflows. But at the pace of research we've seen over the last yr, what I might have thought would take 10 yrs should take 1-2
It’s not just artists: in time, the two sides of the brain within a studio (coders and artists) will converge too. Artists will be able to program game logic through a text based interface with no knowledge of code or visual scripting
Eventually, further automation and abstraction will take shape, allowing anyone through natural language to be able to generate a unique scene to step into
Something as beautiful as this:
Something as beautiful as this:
I don't believe GenAI will eliminate professional game dev or 3D artist jobs:
1/ There is an ever growing demand for high quality 3D content in games and beyond (see: metaverse)
2/ The shift from printing press -> Typewriter -> Word -> GDocs didn't make everyone a novelist
1/ There is an ever growing demand for high quality 3D content in games and beyond (see: metaverse)
2/ The shift from printing press -> Typewriter -> Word -> GDocs didn't make everyone a novelist