top of page
00179-496819074_edited.jpg

HERZOG

RESEARCH

Character creation

Character consistency

Reference sheets

Style Variation

RESEARCH

Stylised skin rendering

Biological form mutation

Lighting control

CREATURE

FANTASY

RESEARCH

Stylised world detail

Distance consistency

Lighting control

Video consistency

RESEARCH

Stylised world detail

Environment mutation

Scalable environment damage

Video consistency

SCI-FI

IMPORTANT INFO
 
I am currently exploring generative rendering workflows, with an initial focus on concept development and game mock-ups.
 
This research is driven by a few personal questions:

 

  • Can artists steer the tools accurately?

  • Can we avoid cloning others' work?

  • Can it uncover new creative directions?

  • Could style lora models be a regulated format?

  • How might style licenses work?
     

These tools aren't going away; they’re going to evolve and be adopted across creative industries. Artists should be the ones exploring this technology and, hopefully, defining how and where it's used. Maybe we can even validate its artistic potential and help define an acceptable legal framework? But if we bury our heads and avoid what's developing, unregulated mediocrity could overwhelm everything, and artists will be devalued even more.
 
I want to be sensitive to the complex and passionate discussions surrounding these technologies. I always prioritize ethical data practices where possible. I work with offline AI models, bespoke training, and open-source tools, and I don’t support unregulated data scraping.
 
Much of this work is purely process and pipeline focused. It's early days, and wherever possible, I directly steer style, composition, lighting, and details using a combination of handmade Photoshop layouts, ControlNet tools, prompts, and custom style models built from internal references I’ve rendered or photographic data I’ve collected. I also take care to avoid directly copying or referencing other artists' work.

bottom of page