2🐕Lighting Class

Lighting language is one of the areas where the exact tag really matters - "dramatic lighting" and "rim lights" and "Rembrandt Lighting" all point at genuinely different setups, and if you don't already know the vocabulary you'll default to vague words the model half-ignores. EG_TSCDS_DG ("Lighting Class") is the smallest of the ergouzi "Fixed" category pickers - just two dropdowns - but it's specifically built around getting that vocabulary right.

How it works

Two categories, one node. Light perception covers the overall lighting setup - studio lighting, film lighting, dramatic lighting, Volumetric lighting, Soft illumination, rim lights, Back lighting, Split Lighting, Rembrandt Lighting. lighting covers specific optical/lighting effects layered on top - lens flare, backlighting, sparkle, bokeh, caustics, chiaroscuro, cinematic lighting, ray tracing, god rays, glowing. The distinction is roughly "what kind of light setup is this" versus "what visual light phenomena show up in the shot" - you can combine one from each, e.g. Rembrandt Lighting (setup) plus bokeh (effect), and get both in your output string.

The inputs and outputs that matter

Both categories default to "None" and carry their own weight slider (0.1–2.0, default 1). The practical range for keeping tag emphasis useful rather than distorting is roughly 0.5–1.5 - lighting tags in particular tend to visibly warp an image if pushed past that (over-weighted "god rays" or "lens flare" can blow out a whole frame). random (default no) plus seed roll a random pick across both categories instead of your manual selections. Output is a single prompt (STRING) to concatenate with the rest of your prompt.

How to install it

ComfyUI Manager: search Comfyui-ergouzi-Nodes. Manual install:

cd ComfyUI/custom_nodes
git clone https://github.com/11dogzi/Comfyui-ergouzi-Nodes.git

Restart. Same standing note as the rest of this pack: the author's README says this English repo is no longer updated, with active work continuing on a Chinese-named sibling pack.

Common issues & troubleshooting

Lighting tag doesn't visibly change anything. Lighting is one of the more checkpoint-dependent tag categories - a model heavily fine-tuned on studio photography will respond strongly to "Rembrandt Lighting," while a general-purpose or anime-tuned checkpoint may barely register it. If a specific lighting tag isn't landing, try a couple of adjacent options from the same dropdown rather than assuming the tag is broken.

Weighted lighting tags mean nothing on LLM-encoded models. As with the rest of the pack's prompt-word nodes, (rim lights:1.4) only functions as emphasis on CLIP-based checkpoints. On Z-Image, Flux 2, Anima, or Krea 2, that syntax is literal text passed through unweighted - describe the lighting in plain language instead on those models.

Combining too many lighting effects muddies the shot. Because lighting and Light perception both feed the same output string, stacking a strong pick from each (say, Volumetric lighting plus god rays plus lens flare) tends to fight itself rather than compound - pick one dominant light setup and at most one accent effect.