- This topic has 21 replies, 6 voices, and was last updated 1 year, 6 months ago by
.
-
Posts
-
@eelgoo Hello eelgoo nice to hear from you. 🙂 I also totally agree with you.
Its another level separating the art from the artist. Its like saying since I didn't create the paint and brush myself its not a valid piece of art.
In some cases you have less control of what the AI spits out. Keep entering random text till it looks like something you like rather than actually controlling the figures.
Few reasons I use Daz.
One is some of the content is past my skill to create.
Two some of the content Im to lazy to create.
Three it makes creating content easier.All of these reason can and probably do apply to AI. However the difference is the brush is still in my hand. Letting AI do it I am handing the brush to a RNG AI and letting it think and act for me. Would you take credit for a commissioned piece of art you told someone to make?
Its a form of plagiarism. If you asked me whose figure and clothing I put in my render I could tell most of the time. If I keep random text into a AI I couldn't tell you what artist the computer took from.
I have seen some really amazing AI images bit I just cant accept it as some others have. Grab a few cans of paint randomly pour it on a canvas and call it a day.
What you are saying, might be true for some AI image creators like Midjourney , where you just prompt, get a cool image, but you have very little actual control over the details.
Stable Diffusion, on the other hand, is a totally different beast. It has been designed to give the user the most freedom, you can download extra clothing, characters, poses, styles, etc, and you can even use extensions like ControlNet where you can to create more precise compositions (you can control poses and composition with it).
Certainly, there are some limitations with it, but it's advancing really fast (yesterday we just got the next gen of stable diffusion to make even bigger images, with more detail, and faster).
Its a form of plagiarism. If you asked me whose figure and clothing I put in my render I could tell most of the time. If I keep random text into a AI I couldn’t tell you what artist the computer took from.
That's the nonsense that some artists typically say, so please don't let them poison you with it. The figure and clothing you get is an amalgamation of many different images. Not to mention that that is exactly how human imagination works (by putting various elements together to create something new) (after all AI is modeled after the human mind, in a way), so if it is a form of plagiarism, then every artist on the planet is a plagiarist by virtue of doing the exact same thing.
So I watched this video of a guy promoting the newer version of this program. He selected an image of an astronaut and a cat. Clicked a button and made a cat astronaut. While it looked nice and seems like a near limitless amount of fun. I really can't become attached to what would take me all of 10 seconds. If I were to kitbash 10 images and alter it in PS I guess I'd feel more of a connection.
@d0gg0d I agree that if the creative process is too simple, then there is no feeling of accomplishment and the whole thing becomes pointless.
Now, the video in question is just an aspect of AI image creation, the so called "img2img" in which an existing photo is transformed into something different. Of course, you could take one of your renders and do that too...it would just give you an extra dimension to your creativity.
As for the kitbash...yes, you could do that too, and then use the img2img function to create something new too.
Overall, this is just a tool, some people are going to use it with more complexity, more addons, and more creativity, creating something truly amazing, while others are just going to use it in a more direct and simple way, creating something more "typical" and less appealing content.
And in terms of complexity, there can be a lot of it too. I am currently assessing which of the 20 samplers (similar to a render engine) is better, learning about artistic style of various authors (to recreate their style).
On top of that, you can also use poses and expressions from DAZ characters to create the composition that you want (although this option is not working yet with the new model, who has been out for less than one week, but see my previous comments in this thread).
It's an interesting form of creation and I may mess with it one day. I guess the thing you said that stood out to me is the AI and you are learning the style of other artists. I've no doubt that over the years people are affected by the art styles they see but at some point they should probably develop their own style. When the AI says to itself I have referenced 1 billion images and I personally like these 5 artists and here is my own style based off my preferences. That would be something.
If I select a couple Giger images and mash them together with some Victoria Secrets models does that make it my style?
Maybe it's not different than mashing a bunch of Daz products together but somehow it "feels" different.
- You must be logged in to reply to this topic.
