Layer Breakdown: The Final Push

As many of you know, I've been playing a lot recently with AI generated artwork, especially creating robot designs. However, one thing I've been unable to achieve in midjourney (the AI I've been currently using) is a 3/4 view of the robot designs. All of them end up being frontal view only. This is likely because the images the AI trained on either haven't been 3/4 view, or maybe they have but that particular attribute hasn't been labeled, and so the AI hasn't learned what a 3/4 perspective is. While I'm sure this will be fixed in the future, I've been playing with giving the robots depth by projecting the artwork onto simple 3d geometry. So this tutorial discusses this technique which I used to make my painting "The Final Push".
For more info and other lessons, please visit www.neilblevins.com/art_lesson...

[@JasonKey3D]

Thanks for sharing 🙏 Much more interesting than everyone's posts of straight MJ output

[@ArtOfSoulburn]

Thanks Jason. Ya, the "straight from the ai" posts got old quick, I want to use it as a tool, not as a replacement :)

[@kuzwk]

Cool and inspiring concept. Thanks so much for sharing your work flow . Do you test some AI algorithms? I try Disco Diffusion, but also heard that Dall-E can be good choices.

[@ArtOfSoulburn]

Thanks kurwk! So I've tried a bunch of different AIs, disco diffusion has a number of extra features, but I've found the interface to be more clunky. Midjourney has fewer options but is easier to use. And I haven't been invited into the dall-e 2 beta yet. But the moment I do get invited in (or the software is released publicly), I'll certainly be running a ton of tests in it to see the similarities and differences.

[@kuzwk]

@@ArtOfSoulburn that's awesome I can't wait to see your tests of this tools. Good luck men!

[@petermgruhn]

I might guess that the alg is based on 2d transformations. It's a flat collage maker inherently.

[@ArtOfSoulburn]

Ya, one bothersome thing I've run into is when experts talk about the algorithms, they either go way too technical or way too vague, and so I feel my knowledge is really incomplete in this area. I do know that these algorithms know nothing about perspective, they're just mimicking the perspective they've observed in the images they were trained on. So it's still possible I suppose they could be trained to know more about perspective. But that also may require a whole new algorithm mixed in with the current diffusion algorithms that all of these software are using. In the meantime, I'll use the manual method :)

[@juanmilanese]

been playing with midjourney. Amazing piece of... software? tech? wathever, is amazing. The workflows you are developing here in your latest videos are super relevant.

[@ArtOfSoulburn]

Thank you Juan! Glad you're finding these relevant! And ya, I am kind of viewing this software as a subcontractor, I give it a design brief and it gives me art, and then I can do whatever I want with the art :)