Just stop after the train scene. It was a decent matrix sequel until they started trying to be a matrix sequel.
The Heiress
Sadly lost the prompt due to editing but model is most likely "Asianface collection and 2.5d remix: v2.0" (looks like it's been deleted)
Fingers can be tricky for ai generation to get right but getting it to hold a weapon correctly is pretty much impossible. Maybe there's a lora for it.
Danger Mouse
amazing detail, best quality, high quality, highres, absurdres, masterpiece, digital art, 1girl, illustration, leotard, long sleeve, erotic, skimpy, lewd, long legs, head accessory, simple background, attack Negative prompt: lowres, worst quality, bad quality, badhandv4, easynegative, leather, latex, rubber, see-through, game cg Steps: 28, Sampler: DPM++ 2M Karras, CFG scale: 5, Seed: 2499687385, Size: 472x736, Model hash: f6957e654e, Model: 7th_anime_v3_C, Denoising strength: 0.5, Clip skip: 2, Hires upscale: 2, Hires steps: 14, Hires upscaler: 4x-UltraSharp
Looks great! Just waiting for ANYTHING to update so they'll support SDXL, I know ComfyUI supports it now but I can wait a bit.
Even Denuvo cracked games can't be compared because AFAIK the latest cracks don't remove Denuvo, they just emulate it.
How to get consistent perspective lines?
Is there any tip to getting more consistent continuing lines when there's some subject or object in front of it? Like for example when theres a fence behind the subject and the height of the fence is completely different on either side of the subject. This happens all the time with windows, railings, seats, and any sort of straight line in the background.
I've thought if one solution is to pre-generate the background partially before adding a subject or something like that.
After refining to 2048px with img2img I usually send it to extras and upscale to 4096px with 4x-UltraSharp but for the most natural looking result I will just scale the 2048px image up in photoshop and then overlay the UltraSharp output somewhere between 40-70% opacity. But then it also depends on the style of the image, for some images that I really want to look their best I'll do both UltraSharp and R-ESRGAN-x4+_anime_6B and then mask out the areas I find to look best between the two in order to get the fewest artifacts. Maximum sharpness isn't always the best, some mild blur and even film grain/noise can do a lot to make an image look really natural and beautiful.
I was late to the party on Captain Marvel and was expecting to hate it but came away decently satisfied with it and I think it's in large part due to the cinematography. It wasn't anything mind blowing but I was surprised to see a more indie-like aesthetically pleasing feel to a lot of the shots in it.