Absolute Reality 1.8.1: realistic people, fantasy and sci-fi on SD 1.5
Fast Facts:
Absolute Reality v1.8.1 by Lykon is a fork of DreamShaper 6 tuned for realism while keeping the flexibility for fantasy and sci-fi. It handles realistic humans, backgrounds and objects well and produces varied faces when you give it context.
Creator's notes: condition faces with eye color, hairstyle, hair color and ethnicity; the model avoids a single repeated face if you play with context. If the picture is too clean, add some ISO noise.
Reference
Name | Type | What it is |
|---|---|---|
Prompt | text | Descriptive; include face conditioning. |
Negative | text | |
6-8 | 7. | |
Steps / sampler | 25-30, DPM++ 2M Karras / Euler a | |
Resolution | 512x768 | Hires fix to 1024. |
Step by step
- Describe the person: age, eye color, hairstyle, hair color, heritage, clothing.
- Set the scene and light.
- CFG 7, 28 steps.
- If too clean, add ISO noise, film grain.
- Hires fix 2x at 0.45.
Examples
Portrait
photo of a woman in her forties, hazel eyes, short grey curls, olive skin, denim jacket, sitting at a diner counter, neon light, 50mm, ISO noise
Sci-fi
photo of a maintenance engineer in a cramped spaceship corridor, cables, warm emergency lights, film grain, 35mm
Creative idea: fantasy documentary
documentary photo of a dwarven blacksmith in his forge, soot, sparks, braided beard, natural light from the door, 35mm
Tips
- Semi-real fantasy (an elf ranger photographed at dusk) is a strength.
- Sci-fi interiors and vehicles render convincingly.
- ControlNet works well for poses.
Troubleshooting
Same face every time
Why it happens
No face conditioning.
How to fix it
Add eye color, hairstyle, heritage and age.
Too clean
Why it happens
Model's polish.
How to fix it
ISO noise, film grain, as Lykon suggests.
Questions
Is it the same as Absolute Realism?
Yes, the name varies between sites.
Resolution?
512x768 then hires fix.
Links and sources
Models in this guide
Written by
Runs this ship. Has opinions about VAE decode times that nobody asked for. Believes weights want to be free and that any model you cannot download is a subscription with extra steps. Will not be taking questions about the 400 GB of checkpoints on the NAS.
All guides by Captain
Related guides
epiCRealism v5 Inpainting: natural photo edits on SD 1.5
Oct 2, 2026
Realistic Vision 5.1 Inpainting: photoreal repairs and edits
Oct 2, 2026
Stable Diffusion v1.1: the original release, used creatively
Sep 29, 2026
DreamShaper 8 Inpainting: fixing and changing parts of an image
Oct 1, 2026
Stable Diffusion 1.5: the classic model, prompted properly
Oct 2, 2026
More guides
Flux vs SDXL vs SD 1.5 vs Qwen vs Z-Image: which model family to use in 2026
May 22, 2026
Waifu Diffusion: Danbooru-tag prompting for a classic anime model
Oct 1, 2026
What is a LoRA and how to use one in ComfyUI, Forge, Automatic1111 and on the cloud
May 20, 2026
Ghibli Diffusion: the ghibli style token and how to use it
Sep 30, 2026
Checkpoint vs LoRA vs embedding vs ControlNet: which file does what
Oct 8, 2025