RetroGlamourAI

I've written an article describing my upscaling workflow. Additionally, I have revised my style guide and written an article giving a technical explanation of some of the setup choices I've made. All of these (and most previous) articles can be be found on my Civitai page.

Recently, a user called PromptGeek has published a fantastic guide on how to prompt photorealistic images. Experimenting with his prompts has led to a real breakthrough for me, and not just for realistic images, but for my regular style as well. Here I share a bit about what I have learned. He does not recommend any specific negative prompts (although he uses a few in his YouTube video, presumably those recommended for each model). With the right mix of negative embeddings, the results can be improved even further. The positive prompt structure he recommends is: [STYLE OF PHOTO] photo of a [SUBJECT], [IMPORTANT FEATURE], [MORE DETAILS], POSE OR ACTION], [FRAMING], [SETTING/BACKGROUND], [LIGHTING], [CAMERA ANGLE], [CAMERA PROPERTIES],in style of [PHOTOGRAPHER] This is similar to, but more elaborate than, the structure I have recommended in my style guide: [modifier], [type of image], [subject], [additional details], [overall mood] While I almost always leave "pose", "action"