Nice video - those images are really interesting. (I would have looked sooner but was streaming a radio programme at the time and my laptop can't do both at once!)
ETA: I tried feeding those terms into Nightcafe and the result was rather less impressive. Maybe it could be worked up to something better with a few more modifiers and a bit more time.
Nighcafe - 3D render landscape Bryce - qjTqYE0u5GMZ9ib29rdM.jpg
So is ai the make art button that has been talked about for years? Is it as satisfying as creating art work in other ways? Who owns the intellectual property rights to ai? What is the ethics involved in creating art work using this method?
ai raises more questions for me than I have answers...
@WendyLuvsCats, great images you seem to have a knack for it.
Questions aside, a friend recently posted an image from ai which after a long dry spell art wise has inspired me to create a morph dialled character.
Midjourney faces can be a bit off at times, sooooo, I have a Faceswaplite app but sadly the only human faces in this house are my ugly mug, not really much of an improvement
Mystery: Thank you very much. I like it myself, I'm not just sure how good of a job it is. ;)
Melanie: No, it was a separate render of a nude figure that I merged with layer masks into the Midjourney render. Pretty much everything about the figure itself (the skin, legs and arms, back cutout, and pipe) were from the Studio render. Lots of other misc postwork, but the majority of the background and about 80% of the dress was Midjourney. Thank you!
Midjourney faces can be a bit off at times, sooooo, I have a Faceswaplite app but sadly the only human faces in this house are my ugly mug, not really much of an improvement
I love it, Wendy! Or should I say Lady LuvsCatz? Haha that even sounds legit.
Does that mean that you could use that face swapping software with a suitable render? You may have just solved the biggest problem MJ has...
Melanie: No, it was a separate render of a nude figure that I merged with layer masks into the Midjourney render. Pretty much everything about the figure itself (the skin, legs and arms, back cutout, and pipe) were from the Studio render. Lots of other misc postwork, but the majority of the background and about 80% of the dress was Midjourney. Thank you!
Woah, I assumed the dress was D-Forced. Can you guys imagine, like Dr. Károly Zsolnai-Fehér always says, what this is going to be like "two papers from now"? It's going to generate textured 3-D meshes before we know it.
Melanie: No, it was a separate render of a nude figure that I merged with layer masks into the Midjourney render. Pretty much everything about the figure itself (the skin, legs and arms, back cutout, and pipe) were from the Studio render. Lots of other misc postwork, but the majority of the background and about 80% of the dress was Midjourney. Thank you!
Wow, sounds like quite a bit of work, but nice job!
Midjourney faces can be a bit off at times, sooooo, I have a Faceswaplite app but sadly the only human faces in this house are my ugly mug, not really much of an improvement
I love it, Wendy! Or should I say Lady LuvsCatz? Haha that even sounds legit.
Does that mean that you could use that face swapping software with a suitable render? You may have just solved the biggest problem MJ has...
yes works with any face including a friend or a doll in camera shot (swaps them and can do video paid version which I won't get)
It let me use the video feature for a limited time and I have this doll with my face as a video somewhere (cannot find it on Facebook)
there is another one called FaceApp many are using in FB Midjourney group with Artbreeder faces
The bigger question is going to be who actually owns the copyright to any content the AI generates. After all, how did the AI "learn" to create these things in the first place? They had to feed the AI millions of images, and a good number of those are likely copyrighted images. I can see that different countries might take different stances on this subject, and that some will create copyright laws intended to protect human artists.
Some countries may pass laws that state you cannot copyright AI created images at all, and that would certainly hinder those with commercial plans. Others might decide that the company behind the AI owns the copyright, which I think would hinder almost anyone using it for even personal art.
Some countries may pass other laws restricting what usage AI generated works can have.
What I see AI doing is generally taking a big chunk of stock image sales. Stock image sites already take a portion of potential revenue from commission artists around the world, as any company or person looking for a cheap image can turn to these sites. This business practice is not new, and even predates the internet. IMO the AI stuff will be used as another source of cheap stock images. At least as long as they are legally able to do so.
I don't see AI generated works really going beyond this, because as I described I fully expect law makers to tackle the subject. If not in the US, certainly in other places. Also, don't expect AI to be a free ride. As these apps get better, they are going to want money for them. The popular Face apps already require paid subscriptions for most of their features. It is a given that these other AI image generators will as well.
Wel they just added another creation method to Nightcafe Creator called "Stable" which can generate people and animals much better than their other methods. I tried it out this evening, and for the first time feel kind of guilty doing this - it seems like more of a cheat than before.
Wel they just added another creation method to Nightcafe Creator called "Stable" which can generate people and animals much better than their other methods. I tried it out this evening, and for the first time feel kind of guilty doing this - it seems like more of a cheat than before.
The biggest problem I see with Stability, Dall-E, etc is the inability to dictate framing and, also, the characters are never consistent.
Also, frankly, to me every image looks the same. It's devoid of style.
The bigger question is going to be who actually owns the copyright to any content the AI generates. After all, how did the AI "learn" to create these things in the first place? They had to feed the AI millions of images, and a good number of those are likely copyrighted images. I can see that different countries might take different stances on this subject, and that some will create copyright laws intended to protect human artists.
Some countries may pass laws that state you cannot copyright AI created images at all, and that would certainly hinder those with commercial plans. Others might decide that the company behind the AI owns the copyright, which I think would hinder almost anyone using it for even personal art.
Some countries may pass other laws restricting what usage AI generated works can have.
What I see AI doing is generally taking a big chunk of stock image sales. Stock image sites already take a portion of potential revenue from commission artists around the world, as any company or person looking for a cheap image can turn to these sites. This business practice is not new, and even predates the internet. IMO the AI stuff will be used as another source of cheap stock images. At least as long as they are legally able to do so.
I don't see AI generated works really going beyond this, because as I described I fully expect law makers to tackle the subject. If not in the US, certainly in other places. Also, don't expect AI to be a free ride. As these apps get better, they are going to want money for them. The popular Face apps already require paid subscriptions for most of their features. It is a given that these other AI image generators will as well.
This is my concern also...who will own the copyright?
Comments
Stable Diffusion handled the prompt pretty well.
these made my jaw drop
How did you do that, Wendy? Is that pure Midjourney or did you feed in images from Bryce/Vue/Maya renders?
Whatever, they are really something!
just the prompts in the image title as they saved with my username
a video with more
Nice video - those images are really interesting. (I would have looked sooner but was streaming a radio programme at the time and my laptop can't do both at once!)
ETA: I tried feeding those terms into Nightcafe and the result was rather less impressive. Maybe it could be worked up to something better with a few more modifiers and a bit more time.
I remember seeing one demonstrated at a "tech"Expo...... I tried it, with the glasses.... got SUCH a headache!!!!!.....
If this is AI, I'll stick with what I've got.
So is ai the make art button that has been talked about for years? Is it as satisfying as creating art work in other ways? Who owns the intellectual property rights to ai? What is the ethics involved in creating art work using this method?
ai raises more questions for me than I have answers...
@WendyLuvsCats, great images you seem to have a knack for it.
Questions aside, a friend recently posted an image from ai which after a long dry spell art wise has inspired me to create a morph dialled character.
My friends image plus my morph...
This is the first image I made combining a DAZ Studio figure render and a Midjourney image. Came out OK I guess.
Those tools quite helpful, especially when I`m having creative blocks about composition - camera placement and color palette .
I dunno, @SnowSultan, I like it. A lot. The figure really looks like it is in the scene.
That's really lovely! Did you put the rendered figure onto the background in a photo-editor, or somehow manage to get Midjourney to incorporate it?
Midjourney faces can be a bit off at times, sooooo, I have a Faceswaplite app but sadly the only human faces in this house are my ugly mug, not really much of an improvement
Mystery: Thank you very much. I like it myself, I'm not just sure how good of a job it is. ;)
Melanie: No, it was a separate render of a nude figure that I merged with layer masks into the Midjourney render. Pretty much everything about the figure itself (the skin, legs and arms, back cutout, and pipe) were from the Studio render. Lots of other misc postwork, but the majority of the background and about 80% of the dress was Midjourney. Thank you!
I love it, Wendy! Or should I say Lady LuvsCatz? Haha that even sounds legit.
Does that mean that you could use that face swapping software with a suitable render? You may have just solved the biggest problem MJ has...
Woah, I assumed the dress was D-Forced. Can you guys imagine, like Dr. Károly Zsolnai-Fehér always says, what this is going to be like "two papers from now"? It's going to generate textured 3-D meshes before we know it.
Wow, sounds like quite a bit of work, but nice job!
yes works with any face including a friend or a doll in camera shot (swaps them and can do video paid version which I won't get)
It let me use the video feature for a limited time and I have this doll with my face as a video somewhere (cannot find it on Facebook)
there is another one called FaceApp many are using in FB Midjourney group with Artbreeder faces
mine on my iPad, no PC versions AFAIK
funny ones out there in general
https://www.facebook.com/uniladmag/videos/5153977077958643
https://reaction.life/the-approaching-tsunami-of-addictive-ai-created-content-will-overwhelm-us/
The approaching tsunami of addictive AI-created content will overwhelm us
AI tailoring TV programs to the viewer
Wel they just added another creation method to Nightcafe Creator called "Stable" which can generate people and animals much better than their other methods. I tried it out this evening, and for the first time feel kind of guilty doing this - it seems like more of a cheat than before.
The biggest problem I see with Stability, Dall-E, etc is the inability to dictate framing and, also, the characters are never consistent.
Also, frankly, to me every image looks the same. It's devoid of style.
This is my concern also...who will own the copyright?
my go on Nightcafe today
handsome elven lord with long blond page hair framing face muscular body open brocade robe revealing many neck chains of gold 8K resolution
prefer Midjourney result
catz
Stable Diffusion GUI
https://grisk.itch.io/stable-diffusion-gui?fbclid=IwAR34DKVXloA9z6T8DXpioks9ERqgPB6AmUWKfs7MMUYt_xNsNuSzlsf2hQI
with the 2080Ti I can only do 512x512 but its free an exe and on my PC (sadly for my Mac user friends only PC)
Your Linux friends are apparently equally out of luck :(
awww but you can probably run it using Notebook Anaconda or whatever I am too dumb
and yes uses all the VRAM, it's a resource hog
Heavy postwork between AI generated elements ( Clothes + Headwears ) with our G8Female
Within 5-7 years, we may see Disney/Pixar quality CG movies created by AI.