I see people posting negative comments towards you and saying they will unsubscribe, but your channel is all about reporting what new tools and developments there have been, whether people like those tools or not. Nobody said they would unsubscribe from you when Blender dropped the game development tools. The results here are surprisingly impressive, and the fact that it's Tencent that has released it makes it important industry news. Of course you need to let people know about it.
7:08 "and ahh.. Speaking about Zbrush, Zbrush do not import GLB files, so you definitely need to use blender and then export that as Obj or Fbx" I believe you may have missed it. listen to it again.
Questions for Blender (4.3.2) 1. What are good settings for the voxel remeshing? 2. What's a good merge distance for the vertices? 3. Should I keep the same number of faces as I exported from the AI?
Is there a solution to use more images with additional angles or details to feed AI model with more information and keep generated modes closer to our needs/designs?
@@ArturJaskolski Exactly what should have been done, there is a workflow using ComfyUI that does just that! search it up, but whatever they do and get advanced it needs cleaning up.
this is exactly what I want. I don't understand why no one is focusing on this. everyone advertises stuff like "generate using only one image". but I have 7-8 images that gives a better idea about different angles and there is absolutely no need to do the guess work. and yet for some reason after a few months, something new comes up, it is again for JUST ONE IMAGE.
I found a photo of a dog and tried it. After re-meshing I had a very acceptable base mesh to play with. This is insanely useful. Shame about the territory restriction in the license though. Seems odd.
Still waiting for something that allows multiple images for reference for multiple angles like top bottom, back, sides, etc rather than just AI doing the guessing work.
hey, I'm not a 3d artist but this type of AI will shorten the base creation model progress, it's pretty amazing but I believe it's just for basic work, you still have to repolish, maybe still have to repo at some point, put on shaders and stuff to make it compelling to audience. And this is free, just use it as tool and make a boost.
It does a good job for distant assets as of now ...topology is better than tripo or Rodin and super fast but still not qualified to be hero assets u less yearning for a style
As someone who have been modeling since 2009 I welcome this tools. I have been wondering, is there a tool that can take multiple references? I don’t want the AI to guess the back of the model. When it comes to remeshing, have you tried simply converting the triangle to squads instead, after merging by distance? And lastly, is there an AI tool that can generate all kind of textures that you know of? Metallic, normals, ambient occlusion, cavity, etc. Answers to any of these question will help me greatly, as I’m working on making my dream game.
still waiting for this to be more than simple cartoon characters and simple things in gen...this is easy low hanging fruit dev so far...will they get to more complexity...im waiting to see.
Chinese AI is dominating! The only real competitor with a chance to keep up is Google. While Google may never be open-source, it will remain completely free.
@ShankarSivarajan Do you even know about Google Labs? Even the best Video generation model comes from google, number one in the world right now, all free, no workaround, hassle to setup, plug and play! Speaking of free AI tools, Flux was initially promoted as open-source, but now they’ve released Flux 1.1, which requires payment. The point is, consumer-grade AI doesn’t need massive models unless you’re running a large-scale project, and that will always come with costs anyway. When it comes to AI accessibility, Google is more than generous in providing access to powerful tools. Go educate yourself on these matters, it’s worth it.
Tencent gives you a free tool to generate free 3d models, you get a free tool, tencent gets tons of free 3d models that they can freely use in their videogames, great strategy to fire all those VG assets artists. Not a surprise the style march with their standard vgs.
Well if you are really concerned about it, no that’s still reason to learn. You need artistic sense to judge if these generated assets are good, and you need to match the art direction so you probably need to modify or even recreate based on the generated assets, production ready assets are generally very specific so don’t just practice the tool, also practice the art
There was a similar panic when quartz watches was introduced. There was a legit fear that nobody would buy mechanical watches anymore. And yet, all it did was made mechanical watches even more valuable for having that human touch. 😊
I see people posting negative comments towards you and saying they will unsubscribe, but your channel is all about reporting what new tools and developments there have been, whether people like those tools or not. Nobody said they would unsubscribe from you when Blender dropped the game development tools. The results here are surprisingly impressive, and the fact that it's Tencent that has released it makes it important industry news. Of course you need to let people know about it.
What do you mean zbrush doesnt take obj? I use zbrush extensively and i only work with objs
7:08 "and ahh.. Speaking about Zbrush, Zbrush do not import GLB files, so you definitely need to use blender and then export that as Obj or Fbx"
I believe you may have missed it. listen to it again.
How do I reapply the texture after remeshing in Blender?
Questions for Blender (4.3.2)
1. What are good settings for the voxel remeshing?
2. What's a good merge distance for the vertices?
3. Should I keep the same number of faces as I exported from the AI?
Regardless of the morality of this technology, I must admit that every year the result becomes closer to the required quality of handwork.
Looking for a tool to generate buildings, and to generate also buildings from a video or a scan
Is there a solution to use more images with additional angles or details to feed AI model with more information and keep generated modes closer to our needs/designs?
@@ArturJaskolski Exactly what should have been done, there is a workflow using ComfyUI that does just that! search it up, but whatever they do and get advanced it needs cleaning up.
Look up Rodin. Paid has multiple angles possible
this is exactly what I want.
I don't understand why no one is focusing on this.
everyone advertises stuff like "generate using only one image".
but I have 7-8 images that gives a better idea about different angles and there is absolutely no need to do the guess work.
and yet for some reason after a few months, something new comes up, it is again for JUST ONE IMAGE.
No but if you give a 45 degree angle instead of a flat 90 view it does a better job
I found a photo of a dog and tried it. After re-meshing I had a very acceptable base mesh to play with. This is insanely useful. Shame about the territory restriction in the license though. Seems odd.
Ye
Still waiting for something that allows multiple images for reference for multiple angles like top bottom, back, sides, etc rather than just AI doing the guessing work.
Should be useful for replacing the free models you can find online and bad photogrammetry models
replace bad free photogrammetry models with bad AI models... ok 👁👄👁
"Generate a man standing in front of a row of tanks in an act of protest against a oppressive regime."
I am wondering if there are some devs out there who are about to make a rewrite of that code?
Maybe in china
@diliupg rather Eupean ones because of the licence restrictions here
goodbye to my job
If that’s the quality of work you have done so far, then I’m actually surprised that you have a job after all.
@@stout8529 Burn!. Hey be nice, may have been a foresight comment.
it was already insanely competitive
You can use ai as anyone.
hey, I'm not a 3d artist but this type of AI will shorten the base creation model progress, it's pretty amazing but I believe it's just for basic work, you still have to repolish, maybe still have to repo at some point, put on shaders and stuff to make it compelling to audience. And this is free, just use it as tool and make a boost.
It does a good job for distant assets as of now ...topology is better than tripo or Rodin and super fast but still not qualified to be hero assets u less yearning for a style
Now we can have the 3D version of oshi v-tuber XD.
C:
As someone who have been modeling since 2009 I welcome this tools. I have been wondering, is there a tool that can take multiple references? I don’t want the AI to guess the back of the model.
When it comes to remeshing, have you tried simply converting the triangle to squads instead, after merging by distance?
And lastly, is there an AI tool that can generate all kind of textures that you know of? Metallic, normals, ambient occlusion, cavity, etc.
Answers to any of these question will help me greatly, as I’m working on making my dream game.
This is heaven for game assets creators!
Now all you need to knoiw is programming, and even this is being simplified by AI.
still waiting for this to be more than simple cartoon characters and simple things in gen...this is easy low hanging fruit dev so far...will they get to more complexity...im waiting to see.
Yes. Currently good for children to play around with. 😂
Çok faydalı bir içerik
this kind service is like "3d scan in 2d"
Hehe, Nice.
Having issues getting it to work in blender, please help.
have you tried turning your computer off and turning it back on again?
@@Klaypersonhe should have used AAA batteries instead.
awesome. but I take it back. we still got jobs. it not there yet
Very convenient, at the same, so scary. No denying it anymore, this is the future.
Someone needs to use this AI tool for something serious to show its potential. Maybe I’ll be that someone.
lmao
Chinese AI is dominating! The only real competitor with a chance to keep up is Google. While Google may never be open-source, it will remain completely free.
No, Google requires a Google One subscription for their more advanced models.
@ShankarSivarajan Do you even know about Google Labs? Even the best Video generation model comes from google, number one in the world right now, all free, no workaround, hassle to setup, plug and play!
Speaking of free AI tools, Flux was initially promoted as open-source, but now they’ve released Flux 1.1, which requires payment. The point is, consumer-grade AI doesn’t need massive models unless you’re running a large-scale project, and that will always come with costs anyway.
When it comes to AI accessibility, Google is more than generous in providing access to powerful tools. Go educate yourself on these matters, it’s worth it.
@@ShankarSivarajan Not exactly.. You can use AI Studio to get access for free, to the cutting edge ones. DOnt' need the gemini sub
I never got a good basemesh out of it. The texturing is even worse, and smudgy
Tencent gives you a free tool to generate free 3d models, you get a free tool, tencent gets tons of free 3d models that they can freely use in their videogames, great strategy to fire all those VG assets artists. Not a surprise the style march with their standard vgs.
Subscribed
so legit there's no point for me anymore to learn character modeling? 😢
Gotta come up with a gimmick 😊
no, use the tools to make your skill even better
@thtcaribbeanguy yes gimmick
Well if you are really concerned about it, no that’s still reason to learn.
You need artistic sense to judge if these generated assets are good, and you need to match the art direction so you probably need to modify or even recreate based on the generated assets, production ready assets are generally very specific so don’t just practice the tool, also practice the art
There was a similar panic when quartz watches was introduced. There was a legit fear that nobody would buy mechanical watches anymore. And yet, all it did was made mechanical watches even more valuable for having that human touch. 😊
Hitting your lowest here
wdym?
so you want every body to ignore all the AI advancements and pretend it doesn't exist ? world doesn't work like that buddy
Shut up loser
This is a Massive, Systematic, Structurized Tech War against the WEST...😂😂😂😂
This is a Massive, Systematic, Structurized Tech War against the WEST...😂😂😂😂