For best results use:
refiner: expert_ensemble_refiner
refine_steps=50
lora_scale=0.75
prompt: “A CCTV image of [what you want] in the style of TOK” or “A computer 3d graphic of [what you want] in the style of TOK, information, black background, CCTV”
negative prompt: “monochrome, bad anatomy, smudged faces, smudges, old-time” or “blurry, blurred, bad anatomy, smudged faces, smudges, blur, spider, neuron”