메시 AI를 사용한 3D 모델 생성

메시를 사용하면 텍스트 프롬프트에서 3D 모델을 생성하고, 텍스처를 다듬고, AI 기반 도구를 사용하여 모델에 애니메이션을 적용할 수 있습니다.

### 메시를 사용한 3D 모델 생성: 텍스트 프롬프트에서 애니메이션 장면으로
* 🤖 Meshy는 사용자가 텍스트 프롬프트에서 3D 모델을 만들 수 있는 생성형 AI 플랫폼으로, **사실적, 카툰, 로우 폴리 등 다양한 아트 스타일**을 제공합니다.
* 🎨 사용자는 **네거티브 프롬프트를 추가**하여 모델 출력을 개선하고, 모델을 개선하여 품질을 높이고, 텍스처, 환경 및 메시 복잡도에 대한 설정을 조정할 수 있습니다.
* 💰 메시에서는 **월 200크레딧의 무료 티어**와 **월 16달러부터 시작하는 1,000크레딧의 프로 요금제**를 제공하여 사용자가 생성한 자산을 소유할 수 있도록 합니다.
* 🎭 3D 패키지나 게임 엔진에서 사용할 수 있도록 애니메이션을 다운로드하는 옵션과 함께 **Mixamo와 같은 툴에서 리깅을 통해 모델을 추가로 애니메이션화**할 수 있습니다.
* 🌌 이 프로세스는 Adobe After Effects에서 3D 장면을 제작하는 데까지 확장되어 **모델을 HDR 이미지로 조명**하고 **컴포지션에 통합**할 수 있습니다.
* 🧙 메시의 AI 텍스처링 기능을 사용하면 **텍스트 프롬프트에 따라 세부적인 텍스처를 생성**할 수 있으며, 컨셉 아트를 사용하거나 다양한 아트 스타일 중에서 선택할 수 있는 옵션이 있습니다.
* 🚀 이 플랫폼은 귀여운 다람쥐나 카우보이 모험가와 같은 캐릭터부터 **소행성이나 칼**과 같은 물체까지 다양한 모델을 빠르게 생성할 수 있습니다.
* 🌟 이 동영상 튜토리얼에서는 **3D 모델 생성, 리깅, 애니메이션, 씬 제작을 위한 After Effects와의 통합에 Meshy**를 사용하는 방법을 보여줍니다.

I'm going to show you how you can use generative AI to go from a simple text prompt to an elaborate 3D model in seconds. So I'm going to go through the platform, which is Mesh. And then look at how you could create a simple model, build elaborate scenes, and more traditional digital animation techniques to bring your scene to life. Here we go. Okay, so as I mentioned, the tool I'm using is called Meshy, and there's a link in the video description below. Press Try Meshy for free to get started. Once logged in, you're presented with a very smart UI where you can browse other users' generations that they've shared. You can click on these, check out the text prompts they've used to generate them. You can also remix, copy the seed number, and expand on them from there. I'm going to go over to the AI toolkit, choose Text to 3D, making sure I'm using the new beta released version, which is the best version out so far. And to get started, I'm going to go and add my initial text.

prompt. So I'm just going to add human skull realistic and you could then choose what art style you want to go for from Auto, where it'll choose for you realistic, cartoon, or low poly. And I'm going to press realistic and you could just press generate but I'm actually going to add a few negative prompts to help try and get slightly better output. To get you started, you can press guidelines, scroll down to the bottom, and there's a negative prompt section. I'm just going to copy these and paste those into the negative prompt box and press generate. And then within 20 seconds, it's generated four 3D meshes for us to review. And these are all low quality to start with. Click on this one here, which looks pretty cool, and it's got some here in more funky colors. And you can press the refine tab here or press the settings and press refine. And I found this can take less than a minute to complete. And you see we've got a really cool, pretty realistic 3D model, nice high

polygon count and the texturing isn't bad at all. It's got slightly confused. It looks like there's maybe some of the teeth texture on the head, but I'll show you how you can retexture the model later on. But overall looking really nice. Yeah, very happy with that. In the texture settings, you can go from color to PBR, which will add normals, roughness, and level of reflectiveness or metallic, and you can increase or decrease these. And we now have a slightly metal tool skull. In the environment settings tab, you'll see there is a HDRI image up the strength of that map, and you can rotate it so you can quickly see how your model might react in a final 3D package. Later on, there's also a new mesh settings toggle where you can change from triangles to quad, and you can also reduce the complexity of your mesh. With that done, we can turn on the wireframe again and see we now have quads rather than triangles and far lower polygon count. Press download, and there's various formats available. I'm actually going to

go for a glb file and press downlo download and that's going to cost 20 credits. Okay, and before I do some more model generation to build up an entire scene, I wanted to show off the AI texturing feature within Meshi. Going to AI toolkit, there's AI texturing which is in beta, and clicking the button once inside the AI texturing tool, you can click new project and then upload my 3D file, give it a name, and press create project. You then have the choice of using a text prompt to generate a new texture, or you can use concept art. So perhaps if you've created some imagery using another AI tool, or you've got some photos or illustrations, you could upload these here and generate a new texture, and it will bear those in mind. I'm going to stick with the text prompt approach, and for object, I'm going to say gold skull covered in hieroglyphics. And you can choose the style here, and I'm using ancient style HDR High detail realism negative prompts. Again, I'm just

going to paste in those from the guidelines, and we can select from their various art styles. For this first one, I'm just going to leave it on realistic, and we're going to say use the original UVS and press generate. Okay, and here is the completed retextured model, and it's looking really impressive covered in these gold hieroglyphics, and there's loads of lovely details. There are some areas where the textures got slightly confused, slightly distorted, but you have to remember this is just the beta version of the tool. And like all these AI tools, they're all steadily getting better and better all the time. And I think this retextured cartoon line art style skull is really epic, so much cool detail, love it. Now I'm going to generate a few more 3D models using their text to 3D tool. I'm just going to write cute chipmunk, leave those negative text prompts, change to the cartoon styles, and press generate. And you'll see I've got my four different generations of a cute chipmunk.

and there's a few differences in the mesh in the tail. Choose the one that you think most relevant for what you're looking to create and press refine. And here I have that refined mesh, pretty decent model, but there are some things going wrong with the texture applied to the character. And you will find this with AI 3D model generation. Solutions as it struggles with complex characters. So the way to break this up would be to generate a separate model for the head, the body, the arms, the tail, and piece it together in blender or another 3D package. But still very, very impressive. And this took just a couple of clicks and a quick preview of some other Generations. I had this little spaceship with neon pink blasters generated and this is the refined version. And I produced another set of spaceships, this time using the low poly art style. And this was the final refined version with this very cool low poly sci-fi look, which you can imagine how this in a world with loads of other low poly assets could look really cool and give a

unique style to your animation. I then created this evil Bubblehead figure. If you zoom out, that's going to look really awesome. I had it generate some asteroids, and again, you could very quickly iterate on this, creating different assets to place into your scene, and then update the textures. I generated these different human heads, this guy with a big bushy beard, and this is the refined version, so loads of really nice details in that texture. There's a few areas that could be improved again, like there's a little bit of an eye in the center of his face, but you know, maybe he's got three eyes, but it does just show how far this AI technology is coming in the 3D space. I also generated this Cowboy, slightly more cartoon Cowboy, this intense male Adventurer with some scars on his face, generated a few different exploding planets, this Zelda-inspired big-ass sword using the text prompt big-ass sword, a tree, and the texture of the bark is really impressive, whilst the model kind of loses some of the clarity.

around the leaves, so maybe a more cartoon style would suit the current generations better. This fantastic Unicorn Wizard, and it looks like some sort of Evil My Little Pony character for some future CGI series, a scorpion lizard cat because why not, and a lobster dog. And it's done a great job of combining some characteristics of a dog with a lobster, and a cute little cat looks really impressive with eyes on the back his head. This is one of my favorites. It's a cute Fox character with green eyes, a low-poly princess ninja, and finally this futuristic bear warrior in heavy armor. And I think this is one of my favorites. So in a moment, I'm going to show how you can take one of these characters and try to easily rig it and bring it to life with some stock animation. So I'm going to download the FBX file plus quickly show how you can bring it into After Effects with its new 3D format support and light it with an HDR image generated using Ai. And before I move on to that section of the video, I

wanted to quickly touch on the pricing for Meshi. Um, it's available for free where you get 200 credits per month, and you can have one task in the queue. And any assets are shared publicly with a Creative Commons license. You can get pro access from $16 a month if you pay for a full year. Otherwise, it's $20. And this gives you 1,000 credits each month. You can have 10 things generating at the same time. And you will actually own the assets, the designs that you generate. They then have their maxed tier where you can have 4,000 credits and 20 things generating at the same time. Time. Then as a paid subscriber, you have the option to pay for additional credits. Okay, I'm back to our be character. And I've downloaded a glb file that I can take into After Effects later on. And I've downloaded an fbx file, which I'm going to take over to Mixo to add a character rig and bring him to life. So if you head to mio.com.

and with an Adobe Creative Cloud subscription, you can log in to your account and browse characters that exist already. You can also upload characters and use various bits of pre-made motion capture. I'm going to go ahead and press upload character and select our bare fbx file. I can flip the view round so he's facing forwards, press next, and then position these markers for his chin, wrists, elbows, knees, and groin, and press next. And then it goes through an auto rigging process, and it's completed, and he looks like he's pretty well rigged, and it gives you a little default motion. You can press next, check out our character, you can click animations up here and browse different motions, and it looks like he's working really nicely, he's got some sweet moves, and you can tick in place if you need to, and then once you're happy with your animation, you can press download and download an fpx file with that animation applied, and then you can take it into your preferred 3D package, whether that's Blender, Cinema 4D, Maya, or a game engine.

like Unity or Unreal, and start piecing together a scene. I'm not going to go through in this actual video, as I've covered it in previous ones. Plus, the fact you're watching a video about 3D model generation means you might be familiar with one of those packages already. You can also choose the T-pose to have a static pose of your character, and it will download with that character rig that you can then animate in your own 3D package or apply some AI animation that you might have generated with a tool like Motion or Motion with two O's. I've previously covered that process in these two videos on the channel. And as one last thing, I'm going to take one of our skull models, drop that into After Effects with a background that I'm going to quickly generate, and show how you can light it with some HDR lighting. Now, I've shown this process in different ways over recent videos, but again, I'm going to jump into Zoe depth on Hugging Face, and this is a version made by Sharik Farooq.

you head over to here, press image to 3D, and choose a background image. And I've got this one that I generated using mid-journey. And then tick keep occlusion edges and press submit. And within a couple of seconds, we have this 3D mesh with the depth applied to the scene. I can then download a glb 3D file. I'm then going to head to Blockade Labs where you can create AI-generated 360 images, 3D worlds, and HDR images. And I'm on the page tier, which allows you to download HDR images. And I'm going to add the text empty room in Dungeon dust clouds light shafts greeny blue tones and choose a style realism and press generate. And it's generated this really cool image. And while it's not exactly the same as our mid-journey image, there's enough similar lighting that we can use it for our project today. So I'm going to press download and choose HDR. I've then opened Adobe After Effects, which is part of the Adobe Creative Cloud suite, and I'm

creating a 4K comp at 30 frames per second, and I've already imported one of those skull models that Zoe depth scene and that HDR image from blockade Labs. I'm going to drag down our Zoe depth scene first of all, drop that in, press okay, press R to bring up the rotation properties, flip it 180° on the y-axis. Right click, select new, and press camera, and I'm going to choose a 35 mil camera, and we can move the camera around in this scene, and as long as you stay within certain limitations of the movement, it looks pretty awesome already. Then going to add our skull model, press okay, change the views to four views so we can see what we're doing, and our skull is positioned over here. I'm just going to drag that over above our 3D mesh. You can see it going through the mesh here. Press S to bring up the scale properties and reduce the size of our skull. Jump back to one view where we're on the active camera, press R, and rotate that y-axis.

whilst we're here, we can press the stopwatch, set a keyframe at the beginning of the timeline. I'm going to move to the end and rotate our skull. I'm then going to grab our HDR lighting image, drag that down, and then right-click and press new light, and create an environment light. Press okay, and in the environment settings, making sure you've got it available with this button here, going to change from default to that HDR, and that instantly affects the lighting being applied to our skull and our 3D scene. And we can go through and play around with the intensity, the amount of shadow darkness, and it helps tie both those 3D models together within our scene. You can then temporarily turn off that environment light as it helps speed up your preview window and go through and set some keyframes for the camera, plus potentially bring in some stock dust clouds from somewhere like Envato Elements that we can drop into the scene, a blending mode like multiply or screen so they fit within the environment, and adjust the settings, adding some color.

grading until we're really happy with the final result. I'm going to quickly go through and add a few more touches to the scene and then render it out. But essentially, this is the process for generating a 3D model using Meshy and bringing it into Adobe After Effects. I could, of course, spend many hours refining the scene, making it look just right, but for this tutorial, it will do. I hope you've enjoyed this video looking at 3D model generation using Meshy. And as always, please press like, subscribe, and leave any comments. [Music] Cheers.

Create Stunning 3D Models & Textures With AI  - Full AI Animation Tutorial

Welcome to AI Animation, the channel that brings you the latest showcases, news, and software tutorials about AI Animation and image creative tools. Visit https://aianimation.com to showcase your AI video and animation work, find clients and more.

Our new channel is focused on providing you with the best content on AI Animation, including the latest trends, tips and tricks, and expert advice. Whether you’re a beginner or an experienced professional, our videos will help you stay up-to-date with the latest developments in AI Animation. We cover everything from the basics of AI Animation to advanced techniques and software tutorials. So if you’re looking for the best content on AI Animation, be sure to subscribe to our channel today! 😊


AIAnimation

메시 AI로 정교한 3D 모델 만들기

메시 AI를 사용하면 다양한 아트 스타일로 텍스트 프롬프트에서 3D 모델을 생성할 수 있습니다.

이 도구는 생성된 모델을 다듬고 텍스처, 폴리곤 수, 복잡도를 조정할 수 있는 옵션을 제공합니다.

그런 다음 사용자는 프로젝트를 위해 다양한 형식으로 3D 모델을 다운로드할 수 있습니다.

메시에서 AI 텍스처링 및 3D 모델 생성 사용

메시는 베타 버전에서 AI 텍스처링 기능을 제공하여 사용자가 텍스트 프롬프트나 컨셉 아트를 사용하여 새로운 텍스처를 생성할 수 있도록 합니다.

AI 도구는 아직 베타 버전이므로 텍스처가 약간 혼동되거나 왜곡되는 부분이 있을 수 있습니다.

Meshi의 3D 모델 생성 기능을 사용하면 몇 번의 클릭만으로 여러 개의 3D 모델을 쉽게 만들 수 있지만, 복잡한 캐릭터에는 어려움을 겪을 수 있지만 특히 단순한 모양과 디자인에는 인상적인 결과를 제공합니다.

AI로 생성된 3D 캐릭터와 그 응용 분야

AI 기술은 3D 분야에서 크게 발전하여 다양한 고유 캐릭터와 에셋을 생성할 수 있게 되었습니다.

3D 생성 플랫폼인 Meshi는 다양한 월별 크레딧과 생성된 디자인을 소유할 수 있는 옵션으로 무료 및 프로 액세스를 제공합니다.

생성된 3D 파일은 FBX 또는 GLB 형식으로 쉽게 다운로드하여 애니메이션을 추가하고 After Effects와 같은 소프트웨어에 통합할 수 있습니다.

3D 캐릭터 애니메이션에 Adobe Creative Cloud 사용

Adobe Creative Cloud를 구독하면 3D 애니메이션용 캐릭터를 검색하고 업로드할 수 있습니다.

이 플랫폼은 캐릭터에 대한 자동 리깅과 다양한 사전 제작 모션을 제공합니다.

애니메이션을 선택하고 커스터마이즈한 후 파일을 다운로드하여 Blender, Maya, Unity, Unreal 등 원하는 3D 소프트웨어 또는 게임 엔진에서 사용할 수 있습니다.

어도비 애프터 이펙트에서 AI 생성 이미지로 3D 장면 만들기

이 과정에는 배경 이미지에서 깊이가 있는 3D 메시를 생성하고 glb 3D 파일을 다운로드하는 것이 포함됩니다.

Blockade Labs는 스타일에 맞게 커스터마이징하여 다운로드할 수 있는 AI 생성 HDR 이미지를 제공합니다.

Adobe After Effects에서 4K 구도를 만들고 3D 모델을 가져와 장면을 구축한 다음 카메라와 모델을 최종 출력에 맞게 조정합니다.

어도비 애프터 이펙트에서 메시로 3D 모델 만들기

메시를 사용하여 처음에 키프레임을 설정하고 두개골을 회전한 다음 3D 장면의 HDR 조명을 조정합니다.

환경 조명을 추가하면 조명과 그림자 효과를 향상시키는 동시에 카메라 키프레임을 조정할 수 있습니다.

마지막으로 스톡 먼지 구름과 컬러 그레이딩을 통합하면 렌더링하기 전에 3D 모델을 더욱 개선할 수 있습니다.