You should also know that this building was not built by me from scratch! I'm not an architect, sorry xD I actively used a ready-made 3D model and AI in the process — first I found a suitable house model, then asked Chat GPT to add a floor and generally increase the size of the building. I selected the right angle, then manually made the clean line (screenshot 2), base fill, light and shadow, base coloring + textures. Then I asked the neural network to make several variants with different levels of detail of the semi-finished image and took a few successful elements via photobashing — for example, the bas-relief under the roof, more detailed steps, the stone base, part of the lattice on the veranda. Thus, I criminally saved myself some time "D Well, then I played around with lighting, palette, and environment, added more details like cracks, trees, ivy on the walls, added some fog and a sinister atmosphere. I also realized at some point that I was getting three different shots with completely different color schemes, and spent some time "balancing" them so that the shades harmonized a bit better with each other. I also refused trees in the foreground and the sound effect of rumbling so as not to cover the resulting image. I like the result too much, sorry xD

If it was interesting, I can periodically share such processes with you (if, of course, you're not sick of the use of neural networks and ready-made models).