Thanks

This will be fun and inspiring.

For sure

Well this is a very mild example of AI usage

@@Settiis Now you have the tools you need to go WILD

Its the honeymoon period for AI creativity. Seen as collaborative tool right now. Soon it will take over completely. Its inevitable

@@htyzc There's definitely a love affair with it right now. Work with enough though and the current limitations become obvious. I hope to use it as a tool for generating ideas and mutations

Great tip Chris, thank you.

I will try those settings, thanks for this!

It helped that i made it in 3DSMax first, so I knew the steps, i just had to locate the Blender equivalents and muddle through :D

It's getting there, quickly. Would be especially relevant to engines where polycounts are less of an issue, eg. Nanite in UE5

Yeah I want to make a mini for dnd out of an image but man… I’m just not quite up to this level of computer wizardry.

You're welcome

You are a man sir. You will never be seen as a woman lmao. At least you'll make it to the 41% you sussy baka

I love the tips, I'm a Blender noob, so thank you!

+1 remindme

You're welcome Mark

Fractal noise was huge playground in the 2D and 3D abstract art a lot of techno trance or cyber stuff artwork where made this way in the 90s with a lot of Blue and fluorescent purpose color, but that works for anything that look organic.

Nice one Ash!

Cool tip, thanks

What AI is it?

Glad you liked it, thanks

For Midjourney atleast, it is not free, as everybody seems to be pretending. You only get your first 25 images for free. Afterwards you have to pay a subscription fee of 10 or 30$ a month for 200 images and unlimited respectively. I see a lot of video's have this in common, giving the impression that Midjourney is a free tool. And the weird mindset that you shouldn't take other peoples words, because that is "stealing someones creativity skills", but stealing someones actual work (be it drawing, photosgraph, 3D model, ... ) apparently is not.

@@JustAnotherGamer1005 Just use Stable diffusion...

@@JustAnotherGamer1005 25 tries (with a grid of 4 images in each, so 100 images in total assuming you didn't get what you want 25 times) are more than enough to come up with an alien Faberge egg. As mentioned in the info under the video, this will work with any AI image generator such as Disco Diffusion, Stable Diffusion, Wombo etc. Agreed that there is a level of idiocy around the whole "prompt engineering" thing, I guess people want to hold on to some sort of create input when so much is done for you by the AI.

@@gameartexp Indeed, the Faberge egg was a cool idea. And looking at the result, a surrealistic one also 😄. And for experimentation, those 25 are enough. But there are people who seem to assume or portray this as a free tool in your workflow. If it's something you need recurring, it's not free anymore. I've seen a bit of that Stable Diffusion, I think. That one is free, i think? It's nice too see what AI comes up with. And for indie-devs, this is a great tool that can burden a bit of the load.

Nice one Ian!

The input image is key to the detail in the end result, keep refining your prompts, maybe try another AI?

My source assets are here BTW mcpoly.gumroad.com/l/xhzldp

Nice one, hope it opens some doors!

I've tried this www.artstation.com/blogs/petemc/468w/midjourney-to-3d If enough people like your comment I could do a video on it.

@@gameartexp I would love a video on it :)

@@chrisgreenwell3404 keep an eye on here or subscribe ;)

That would be amazing my! I’d love to see that tutorial.

@@thegeekclub8810 Ok I think I know what my next video will be

You're very welcome Lucas, thanks for the kind words.

Thanks for the kind words

Thanks for that!

@@gameartexp Hey, could you advice please, how did you come by with displacement, normal map and othe jpegs?

Thanks John! The sci-fi pistol was for modelling practice really, I made that the old fashioned way using the AI image as reference. I have some ideas on some easier conversions though, subscribe if you'd like to be kept informed

Thanks!

Thanks Patrick

Glad you enjoyed it! Thank you for the kind comments, I hope the tutorial sparks some cool ideas

Glad I could help!

More to come, thank you!

No worries, hope you enjoy it!

Welcome aboard! Finding time is the hard part!

Glad you liked it!

Glad it was helpful!

Great tip! Thanks Darren

Shadermap 4 can bake geometry too, so you could export the egg model back into Shadermap and bake AO. Or, use Blender's baking tools

Glad it was helpful!

Can you explain a bit more please?

Glad it was helpful!

Thanks

Have fun, thank you!

Glad you liked this

Glad you enjoyed it!

Glad you think so, thanks!

Thanks for watching!

Thanks for the kind words, appreciate it👍

@@gameartexp You’re welcome !

Glad it was helpful!

Absolutely, I used that method in some earlier work, just not in Blender. www.artstation.com/artwork/5B2gmz

You're welcome

oof

Thank you

Glad you liked it

@@gameartexp "Liked" isn't even the word for it. LOVED, is a better word. i got a question though, say i wanted to use Zbrush or Zbrush Core "For some people who can't afford normal Zbrush" is there a way to use it for the sculpting part? like could i keep the Normal Map and Texture and stuff on it while in Zbrush? do you know anything about any of that? cause if you could make a Tutorial for how to use Zbrush to help enhance the sculpting part, i think it'd help people as well. i know the widest net has already been thrown out, cause you did a blender Tutorial here, but blender just can't offer as good of sculpting as Zbrush can currently.

@@Glitch-Gremlin Absolutely Zbrush/Mudbox could be used instead. I'd forgotten about Zbrush Core actually, the tools and high poly model management would be better over there. The benefit of Blender is that it has everything under roof and there's less hopping between applications but I'll look into including Zbrush Core next time around. Thanks.

@@gameartexp If you'd like a full version of Zbrush, i can get you a "Maxon One" Zbrush, Cinema 4D, Red Giant and others Sub for ya, it can only be used for Educational but for KZfaq videos im sure that'd work. ArtHeroes gives them out, you just buy one of their tutorials "One of them is only $50" and then ask them to set you up an Educational License sub, and then pay the $10 per 6 months, they say you only need to re-up it every 6 months, so basically $20 a YEAR for all those? great deal for learning them. but i got a sub im not using for it, so i could just give it to you and you can use it if you want, take over the account and stuff. or you could make one with them, too. let me know though if you'd like that.

@@Glitch-Gremlin Thank you so much, i appreciate the offer. I actually have a full Zbrush license so including it in a video should be no problem. Thanks again.

Thank you! Cheers!

Hope you enjoy it!

It might do! Tech is moving so quick this is outdated already

Sorry about that, an issue with the OBS capture settings - I watched it back before posting and it should be possible to work through with the dialog. Thanks.

Absolutely, you can apply the image/displacement to any geometry that has UVs and enough polygon density to support it. Follow this tutorial with the geometry of your choice! Thanks for the kind words

Thanks

No problem!

Thank you, yep Materialize will also get you over the line!

🔥

No problem 👍

You're welcome!

Glad it helped! Go for it, I'm about to upload a character tutorial that may help!

@@gameartexp Excellent. Watched and did the tutorial on the egg, substituted a character generated in Nightcafe. Took 100 credits to figure out how to get a model sheet created. You can use "model sheet" Unreal character metahuman render" as potentially useful prompts. Have Unreal 5.1 installed and it looks similar to Unity thankfully. Resolution is amazing. It will be some time to figure. Unity is hard. Have little coding skill. Use chatgpt to help create code, but it makes mistakes. No way around having to learn some coding. Back to the tutorial.I see the workflow, photopea to shadermap to blender. Put the character texture on the ball. Learned how to move and adjust the uv map and thank you for explaining that so well. I've done tutorials on it, that didn't click, but now see how that works. So far haven't gotten Nightcafe to do back of character. Model sheet shows front view and side views. Tried "model sheet, t-pose, three views, front,back,side " and it's night generating back. Best I've gotten is front which isn't always a straight frontal view, and two side views. Fingers on hand are deformed. Nightcafe has new algo, which is called SDXL. Uses 3 credits, so have been using 5.1 to test results. Have a couple low poly 3d character modesl to see if the texture will map to those. It's going to be hard to figure fingers. Looking forward to seeing your character tutorial.

@@joelstolarski2244 Great progress! As explained in the character video (live now!) coaxing out the right prompt can be really tricky

Great tips, I'm a 3dsmax user so still a Blender noob

:D:D:D

Absolutely, that's the way to go if you are a subscriber to ZBrush and the Substance Suite. This video is primarily aimed at people who don't have access to paid tools

​@@gameartexpthe good Thing is it work with freenstuff krita Can extract a lot of normal Map.

It's my pleasure

Good to know!

Thank you very much!

Well, that's what it is!

Thanks for the tips!

I think everyone is sitting up and noticing now

Glad you think so Daniel, thanks for the comment

Nothing new here in terms of technique, this tutorial is for people who want to take Gen AI a little further, it would likely be new to them

All you need to do art the end is export the file from Blender as STL, then import into your slicer and use whatever your favorites settings are to print in resin or extrusion

@@gameartexp I’ll keep rewatching your video till I get how to make 3d and shape it like you did , amazing amazing work ! Is there also a place where one could understand the key words to use to get something as amazingly complex as yours?

@@gameartexp thank you , is there a way to transfer the shape from original images to skip creating a shape in blender? while it might be not as complex with an egg shape it would near impossible with a miniature lets say due to shape complexity? Thnx

@@MaksAMD KZfaq is as good a place as any to find tutorials on 3D modelling etc. Start with the basics ;)

@@MaksAMD If I understand you correctly, and you want to skip the Blender part, you could achieve something by displacing a plane primitive with the original images, but I doubt you could get much with it. Take a look at this: www.artstation.com/blogs/petemc/468w/midjourney-to-3d

Cool, thank you!

Watch this space, subscribe for updates

@@gameartexp You got it, brother. Keep up the good work.

Ah I didn't mean to exclude Mac users, I don't use one myself! You could try Quixel Mixer - Mac downloads here quixel.com/products/mixer

In the description, there is a link to the scene files on Gumroad

Thanks for the tip Bruce 👍, I'm a noob in Blender

I'm sure you could make a floor plan using AI that could be extruded, it would give you ground floor at least. There are also options for image to image and depth to image in Stable Diffusion which would be worth looking at

DO IT!

cdna.artstation.com/p/assets/images/images/051/889/028/4k/pete-mc-nally-alienegg-3dprint.jpg?1658413526

@@gameartexp awesome

For sure, his influence is everywhere in sci-fi movies

Awesome, thank you!

I guess you checked my other vids!

No worries, thank you for watching

Anything with tendrils (branches in this case) that extrude outwards in different directions would be tricky. You could ceratinly do the trunk like this and then add branches yourself and deform it afterwards

The hemisphere works well I've found, I've created heads etc. in this way. As long as your displacement works reasonably well on a plane, adding to a hemisphere will help pull it into 3D a little better. Some subjects don't translate well with this method.

There are many ways to model, or you could use alpha textures on cards, each AI scene calls for different approaches!

Thanks for the sub!

It would certainly be more difficult

Adding physics collisions is pretty easy in a game engine for a static object. I had a displacement map gradient scrolling along it in early tests that made it pulse slightly. Further work needed to animate it though, like rigging with bones or making blendshapes/morph targets.

To some degree, have you seen this? shunsukesaito.github.io/PIFuHD/

kzfaq.info/get/bejne/qJyYorxhqJOzZ6M.html

Pretty good, not sure it's much improved on 2020 though

Please try it out!

@@gameartexp I will as soon as I can!!!

Yes, check the video info for a browser based alternative

Thanks Fred. The issue is complexity. To make this feasible we need to be able to use the same AI image front and rear on the 3D model.

@@gameartexp Got it, thanks!

@@fredferkel9604 kzfaq.info/get/bejne/qJyYorxhqJOzZ6M.html

Yes is the short answer. In this video I used Photoshop, 3DSMax, Knald (for the normals, height etc.) and Toolbag for rendering kzfaq.info/get/bejne/btuYYNR3tM-bnIU.html Substance would definitely be of use here and Photoshop is useful for the neural filters.

Thanks. It's just shaded mode using Eevee, with a direct light for shadows and bloom enabled in the render properties

Hi there, I use Zbrush and may include it in a future video

I have a video in the works like this kzfaq.info/love/2MSBDFAjMjlUnnY_jn8gnwcommunity

A good 3D artist can model anything, Nvidia are releasing code next week to create 3D models from text prompts

@@gameartexp awesome

@@NitwitsWorld kzfaq.info/get/bejne/qJyYorxhqJOzZ6M.html