From Dull to Dazzling: IC-Light V2 Advanced Lighting Control Workflow
1. Workflow Overview

This is an IC-Light V2-based advanced lighting control workflow designed for:
Dynamic Lighting Synthesis: Precisely adjust light direction, intensity, and color gradients.
Style Transfer: Migrate lighting styles from reference images via IPAdapter.
Cinematic Post-Processing: Includes detail enhancement, color blending, and sharpening.
2. Core Models
Model Name | Function |
---|---|
RealisticVision V4.0 | Main checkpoint for photorealistic generation. |
IC-Light SD15_fc | Lighting control model (download |
IPAdapter PLUS | Style transfer for lighting features (requires plugin installation). |
3. Key Nodes
Node Name | Function | Installation |
---|---|---|
| Loads IC-Light lighting model. | Download |
| Controls light intensity/direction. | Built-in. |
| Advanced style transfer (lighting focus). | Requires |
| Enhances details (e.g., skin textures). | Built-in. |
| Adaptive sharpening. | Via ComfyUI Manager. |
Dependencies:
IC-Light Model: Place
iclight_sd15_fc.safetensors
inmodels/unet
.IPAdapter Model: Download
ip-adapter-plus_sdxl_vit-h.bin
tomodels/ipadapter
.
4. Workflow Structure
Group 1: Lighting Control & Style Transfer
Input: Reference image + lighting parameters (e.g., intensity
0.182
).Output: Latent data with lighting conditions.
Key Nodes:
ICLightConditioning
,IPAdapterAdvanced
.
Group 2: Image Generation
Input: Conditioned Latent + prompts (e.g., "cinematic sunlight").
Output: Initial generated image.
Key Nodes:
KSampler
(usesdpmpp_2m_sde
sampler).
Group 3: Post-Processing
Input: Raw output + reference image.
Output: Final image after detail/blending/sharpening.
Key Nodes:
DetailTransfer
,ImageBlend
,ImageCASharpening+
.
5. Input & Output
Input Parameters:
Image: Reference (e.g., portrait).
Prompts: Positive (e.g., "cinematic sunlight") and negative (e.g., "low quality").
Lighting: Default intensity
0.182
(adjustable viaICLightConditioning
).
Output: Final image (PNG) + comparison view (
Image Comparer
).
6. Notes
VRAM: ≥10GB GPU recommended (IC-Light + IPAdapter are resource-heavy).
Common Errors:
Missing IC-Light model triggers red error alerts.
Overexposure if light intensity >
0.3
.
Optimization:
Use
FreeU_V2
for quality boost (params:1.5, 1.6, 0.9, 0.2
).Reduce
KSampler
steps (e.g., 30 → 20) for faster generation.