Trying to replicate the mystery of horror movies from the silent movie era, 1920s.

Minimax H3:

No particular prompting or triggers are needed (but they might help for setting/clothes, etc). Some examples contain them, because I copied the prompt from other variants.

Flux2 Klein:

Trained with the Rose optimizer as opposed to the more common Adam variants.

Seems to do well for general silent movie era look.

Works decently for "converting" images to said look with:

"change the image to 20s_horror while maintaining the composition"

or leave out the "maintain composition" part for better style adherence.

Hunyuan:

It can be a bit difficult sometimes, if so it helps to give it some hints, like "1920s", "black and white", "silent movie".

Many of the examples required a frustrating amount of re-rolls.

All showcase images/videos contain ComfyUI workflows.

Wan 2.2:

The Low noise lora has absolutely most effect for style. High noise still affects the outcome, but to a much less degree.

For the showcase, I use 25 steps: 20/5. dpm++_sde scheduler in ComfyUI.

Some prompt examples:

"A woman in a white hospital gown is lying on a bed. A man in a dark suit and a woman in a light-colored dress are attending to her. The setting appears to be a hospital room, with large arched windows in the background. The walls have a textured, almost cave-like appearance, adding a sense of isolation or otherworldliness. The image has a dated feel, likely from a 1920s horror film."

"A vampire, standing next to a stone statue. The vampire is dressed in a dark, Victorian-style suit with a high collar and a cape, typical of his iconic portrayal. He is holding a small object in his gloved hand, with a sinister expression on his face. The background features a stone wall with a textured, aged appearance."