Thread: AI-generated Fan Art Megathread - Create and share your character!

Originally Posted by Ara Myrr

I do not.

Your trailer idea is also far to complex to generate, I've cut everything it can't do, here's what's left:



Seriously, you are vastly overestimating AI's ability here.
If you want something like this you'll need dedicated models/artwork for every scene you've envisioned because it's far to specific.
Look through the thread, it's mostly Elves, Humans and Orcs.. why? Because it's not particularly great at anything else.

Tell that to the people who tell me it's simple on that thread.

Thanks for the reply.

Originally Posted by username993720

Do you guys generate videos, as well?

Originally Posted by Ara Myrr

I do not.

Your trailer idea is also far to complex to generate, I've cut everything it can't do, here's what's left:



Seriously, you are vastly overestimating AI's ability here.
If you want something like this you'll need dedicated models/artwork for every scene you've envisioned because it's far to specific.
Look through the thread, it's mostly Elves, Humans and Orcs.. why? Because it's not particularly great at anything else.

It's not impossible though, there's a workflow now that will work though, not extremely well.



The workflow would involve getting the first frame of each scene in the storyboard, using an AI-image generator. Then using AI image-to-video to animate that initial AI-generated frame. With the best models available right now you'll get 15 seconds of footage for each scene.

And on top of this you'll have to produce all of the audio yourself, separate from the video generations. Do that for every scene in your film, edit it together, and you've got it.

The problem is that it's still a very difficult, time consuming, and expensive process.

Originally Posted by shoc

It's not impossible though, there's a workflow now that will work though, not extremely well.



The workflow would involve getting the first frame of each scene in the storyboard, using an AI-image generator. Then using AI image-to-video to animate that initial AI-generated frame. With the best models available right now you'll get 15 seconds of footage for each scene.

And on top of this you'll have to produce all of the audio yourself, separate from the video generations. Do that for every scene in your film, edit it together, and you've got it.

The problem is that it's still a very difficult, time consuming, and expensive process.

For AI, it's pretty flawless. None of those glitches we're used to seeing. If anyone is up to the task, that'd be great.

LFD? LFD!


Female one have too many tendrils, but I working on it.

Ok, this image was bad. My fault. Redoing it..

Arthas met some of his friends to have a skull tea casually (that means: occassionally!)



- - - Updated - - -

Sometimes, Arthas wanted to be more like his friend, Sauron



- - - Updated - - -

Imagine, you had the power of the helmet of domination and the power of the flame of ainur



- - - Updated - - -

I really think about creating a low rank adaptation model with azerothian gnomes.

For some reason, most models have a very conservative understanding of gnomes being the garden guys with pointy hats.

Originally Posted by schmonz

I really think about creating a low rank adaptation model with azerothian gnomes.

For some reason, most models have a very conservative understanding of gnomes being the garden guys with pointy hats.

You'll probably do best just supplementing the word "gnome" with something else.