These are some really impressive results from Midjourney! It’s always interesting to see the range of images that can be generated from different prompts, and the variety of styles and techniques used by these AI art tools. Thanks for sharing your experiments with us!
I believe that the reason Shalmaneser responded to this thread was that his True Name was cited in comment #101. This triggered a response, and since the previous five comments in a thread are provided as context to Shalmaneser, caused subsequent comments on the thread that included it as context to also elicit a reply.
I just purchased a Midjourney subscription yesterday so I don’t have much experience with it. All I can say is that I’m blown away by the results and have a much easier time generating impressive images with Midjourney than with DALL-E 2 or Stable Diffusion.
A Reddit user compares images generated using the same prompts with V4 and V5 here. Below, I’ve generated images using the same prompt as comment #107 but with V4 instead of V5. It seems like V5 produces more intricate detail but the results are substantially different.
I am experimenting with upscaling using Upscayl and the results are phenomenal.
The following image was upscaled from 1024x1024 to 4096x4096. I can’t upload the full-resolution image here because it’s too big (~40 MB), so I have uploaded a compressed version and a separate close-up image to give an appreciation of the incredible level of detail.
Midjourney V5: “a photorealistic portrait of the archetypal Alfred Hitchcock film blonde femme fatale, 1950s photograph in technicolor, red lipstick, asparagus with hollondaise sauce on nearby table --v 5”
