Walk the plank!

Thanks! I was really happy with how it turned out. After generating, I just put it through img2img with SDUpscale (not the ultimate upscaler) using 4xUltraSharp at .30-.35 denoise, it seems to do a good job. I definitely will be trying some others later on!

Thanks!

cinematic photo epic lighting, beautiful landscape with [magical bioluminescent vibrant alien flowers]:1.3 . 35mm photograph, film, bokeh, professional, 4k, highly detailed

Negative prompt: drawing, painting, crayon, sketch, graphite, impressionist, noisy, blurry, soft, deformed, ugly

Steps: 60, Sampler: DPM++ 3M SDE Exponential, CFG scale: 6, Seed: 2499950887, Size: 768x512, Model hash: f61328e842, Model: RealitiesEdgeXL_4, VAE hash: 235745af8d, VAE: sdxl_vae.safetensors, Style Selector Enabled: True, Style Selector Randomize: False, Style Selector Style: Photographic, Version: 1.6.0

Of course! It’s the one I use to make myself phone and PC wallpapers.

professional photography, epic lighting, landscape shot (high resolution, 4k HDR)

Using AbsoluteReality from Civitai, DPM++ 2M Karras @30 Steps, CFG 5.5 (768 x 512 then upscaled)

I was actually just about to give up completely on the seed and try something else altogether, and deleted most of my prompt that was trying to specify colours, shapes, etc. Now that I think of it, I usually have better luck with a simpler prompt.

(professional photography:1), (epic lighting:1), (geographical landscape shot:1), (exotic alien planet) (high resolution, 4k hdr)

Negative prompt: blurry, out of focus, bad drawing, sketch, text, annotation, writing, font, letters, ((people)), ((watermark))

Steps: 30, Sampler: DPM++ 2M Karras, CFG scale: 8.5, Seed: 1575107236, Size: 768x512, Model hash: bfea7e18e2, Model: AbsoluteReality, Version: v1.4.0

Was a lot faster than I was expecting! I think it was probably ~2 hours, and a good amount of luck to find it. If I could wrap my brain around inpainting, or had any kind of actual artistic skill, I could have added the bits of rock/scale that kinda fell away as the “seeds” lifted away, and they had some more variety between them too.

I’m by no means an expert, but happy to help where I can! I use A1111. It’s a very easy to use (and setup) way of accessing Stable Diffusion, gives you loads of control/settings, and is all locally hosted. Here’s the installer (or google A1111, I wouldn’t trust a stranger on the internet either!) https://github.com/EmpireMediaScience/A1111-Web-UI-Installer

I think most of the greenery came in when using SD Upscale afterwards (4x UltraSharp). Sorry, I forgot to add that bit to the Workflow!

Thanks! Here’s the info from A111’s handy Generational Information:

(professional photography:1), (epic lighting:1), (geographical landscape shot:1) (high resolution, depth of field)

Negative prompt: blurry, out of focus, bad drawing, sketch, text, annotation, writing, font, letters, hands, clocks, gauges, gears, ((people))

Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 5, Seed: 3329466699, Size: 768x512, Model hash: bfea7e18e2, Model: AbsoluteReality, Version: v1.4.0

Thanks! It really is. I love the sheer endless possibilities and variations it produces.

Yeah, it’s an AI. I’ve played with a few, but Stable Diffusion is free (depending on how/where you host it) and gives you a lot of control/settings to make some really cool things.