directed by fuzzy logic
Viewing sample resized to 31% of original (view original) Loading...
Description

I do my best to denoise all my images, since AI tends to introduce a lot of noise on what should be perfectly flat colors. I'm quite certain that nobody notices or cares. It's more to appease my own OCD than anything.

  • Comments
  • And they look great! You're still doing wonders with the fully_clothed tag.

    Would you consider doing a set of wallpapers with lovely scenery and some tastefully cute furs like this? (The kind of image a self-conscious person wouldn't mind if someone else noticed it as their phone background.) Sure some of your images already make great backgrounds, but I think going into the process with a wallpaper as the goal could make some nice stuff. Peaceful forests and fantasy ruins, portrait for phones and landscape for desktops, that kind of stuff. Thanks!

  • Reply
  • |
  • 1
  • digitaldigger said:
    And they look great! You're still doing wonders with the fully_clothed tag.

    Would you consider doing a set of wallpapers with lovely scenery and some tastefully cute furs like this? (The kind of image a self-conscious person wouldn't mind if someone else noticed it as their phone background.) Sure some of your images already make great backgrounds, but I think going into the process with a wallpaper as the goal could make some nice stuff. Peaceful forests and fantasy ruins, portrait for phones and landscape for desktops, that kind of stuff. Thanks!

    I'll see what I can do for ya :) Anything in particular that you would like?

  • Reply
  • |
  • 1
  • Oh, I kinda feel presumptuous enough already suggesting the above things, I wouldn't want to impose. But if I may maybe explain my earlier thoughts a bit better; it's always a shame when cropping an existing image to fit a resolution that part of the scene is lost. Sure everyone uses different resolutions but I think the 16:9 ratio, either landscape for desktops or portrait for phones would fit most. Beyond that, scenes you 'can get lost in' like forests, fantasy ruins, nightime cities etc give your eyes something to wander over that arent just the character posing in the scene, so you don't get tired of the image. You have done some cute furs and avians in previous images, I wouldn't mind seeing them relaxing or posing in environments that are just as pretty. Hmm, like, this image has lovely scenery, though the character takes up most of it, so maybe something a bit further away from the character?

    Rambled a bit there, thanks for listening!

    Updated

  • Reply
  • |
  • 0
  • I first was thinking about this with this image, post #11963. Something like: if there was a bit more tree to fill the scene and the character was part of it instead of the only focus. But! I have no idea the process of this AI generation or how hard it is to fill out scenery, so what I am suggesting may not really be feasible. Still I'd love to see how you go with it!

  • Reply
  • |
  • 0
  • digitaldigger said:
    Oh, I kinda feel presumptuous enough already suggesting the above things, I wouldn't want to impose. But if I may maybe explain my earlier thoughts a bit better; it's always a shame when cropping an existing image to fit a resolution that part of the scene is lost. Sure everyone uses different resolutions but I think the 16:9 ratio, either landscape for desktops or portrait for phones would fit most. Beyond that, scenes you 'can get lost in' like forests, fantasy ruins, nightime cities etc give your eyes something to wander over that arent just the character posing in the scene, so you don't get tired of the image. You have done some cute furs and avians in previous images, I wouldn't mind seeing them relaxing or posing in environments that are just as pretty. Hmm, like, this image has lovely scenery, though the character takes up most of it, so maybe something a bit further away from the character?

    Rambled a bit there, thanks for listening!

    Not imposing or presumptuous at all! :) AI art is the sorta thing I like to do as sort of an aimless side activity while doing other things... sorta a "set it and forget it" background task, and then if anything inspiring spawns, I'll spend a couple of hours focusing on that until I get an output that I like. Since I never have any defined direction I start with, it makes it real easy to try out new things as requested.

    What you ask might be a bit of a good challenge, actually. Most furry artwork is focused on the character in question, so the machine is trying to recreate that. Zooming out is easy- it's just infilling around the edges- but the result is that you still have an image where the focus is still on the character. You can shift it left or right all you want, but the character will still appear "central," and often out of scale. My first few attempts have all felt rather lackluster, but I'm always up for a good challenge.

  • Reply
  • |
  • 1
  • Aah! I begin to understand a little. I can see how ai generation would be referencing mostly character-focused art, so blending them with the scenery would be more difficult as its used to propping the character up in front of a background, instead of them being more.. integrated? Still, I am eager to witness your results. Good luck!

    Looking for examples, here is another image that is a bit closer to the kind composition I had in mind: post #19634. With that detail though I wonder how much is ai and how much the 'director' actually draws in on top, and then maybe does another ai pass to flesh-out their additions. (kudos to oldhroft, it must be a lot of work)

    Updated

  • Reply
  • |
  • 0