Thanks, didn't know the 3D generators had gotten this far. Really need to test it to believe it because I see a base mesh and rough texture. Remesh, sculpt more or new details, weight paint, improve texture and even an inexperienced developer could have a model better than anything I could have dreamt of back in school (many years ago) . The ability to regenerate a small part of the texture was also really interesting. Would be cool if we could also give a rough model as input to a 3D generator. Either as inspiration or to ask the AI to improve it. Imagine if we could combine mesh painting and a prompt to tell the AI to improve something in a certain way. Ex, give it a WV, highlight the hood on the car and tell it that you want part of the engine to stick out visibly and the AI gives you X number of modified meshes with a visible engine block.
@@KrullMaestaren Rodin has released a feature similar to what you mentioned. I have yet to experiment with it but it looks like you can give some parameters for shaping the mesh. Then combined an image as a reference.
i can't get cura to open the stl from rodin without upscaling it 10000%. then it won't slice it once resized. all it shows is a shadow. what am i doing wrong?
@@undeadly1103 I have noticed the meshes from Rodin are tiny. I usually scale them up in Blender. I am a noob at 3D Printing so unfortunately don’t have an answer. I have heard there is an extra process to properly convert the mesh so your slicing software will recognize it.
@ keep us updated. I am curious if these types of models translate well to 3D printing. I am hoping to eventually make some figurines of a few characters on the channel.
I want to be optimistic and say these can be used in a supplemental way. Like if you get concept art for a character with a lot of 2d cheats in its design. The AI can generate model if you get stuck
Exactly, I have worked with a couple teams now that had no art direction. With help from several of these tools and a handful of others they were able to get publishing dollars. Now they have full art teams and projects scheduled for release. These tools are great for bringing the idea to life. Not as much for the product you are trying to deliver to market.
Looks like you can. Treat the result as a base mesh. Remesh the exported result and brake it at the same time into jacket, body, hat, hand etc. I would also do the weight painting on my own to make sure that armor as an example doesn't behave as jelly. You then touch up the textures and apply them to the cleaned up mesh. You can also add new details to the mesh and texture while working on them.
@@KrullMaestaren You definitely have the concept down for finalizing your characters. This is a great jumping off point but they do need the skilled hand of a 3D modeler to be fully useable. Another advanced trick I have learned is to create head, hands, feet, and body as separate meshes. Rejoin them later in Blender. This will give better results for the mesh and textures versus trying to take one full swing at a whole character.
@@RyanHarris-ix1lm that is a great idea. Could use an app like Krita that has a Stable Diffusion plugin and draw out a character in real time. That would be a really interesting workflow.
Unfortunately there are no reason to use theese tools on trial - it lacks textures resolution and overall quality, so its difficult to estimate if theese could be helpful in real production, even in the best efforts it lacks complexity, texture accuracy to be compared to human-created stuff, so no reason to pay for tool you cannot use
@@jozinzzzzzbazin I am currently working on update videos for these tools. There has been enough interest and paid subscriptions to improve the output of both meshes and textures for most of these tools. Of course the human element will always be required to fine tune any of the outputs to your standards. Depending on the complexity of your project current outputs can already be put to use in a production.
@@InfiniteColours with the latest updates. I feel like Rodin is still the best. That tool really dials on the base mesh. Some of the sharpest details yet. It also allows for multiple image references. They just added a face restore feature. I use all three tools for different use cases, but for character work. Rodin wins in its current state
@@ArtificiallyInked thanks! In my experience rodin beats meshy yh, havent tried tripo but seems to be very good. Which one u think is the best for image to 3d? Love ur videos so much, thanks for them!
@@InfiniteColours Tripo is great also. There will be a video out this week showcasing their new update. The new update gives Rodin a run for It’s money with Image to 3D. Where Meshy really shines is Text to 3D. I use all 3 on a daily basis
@ They are definitely useful tools. Especially for developers that maybe wizards at coding but noobs at modeling. Great way to get your projects up and moving
Have you tried any of these tools? If so what was your experience?
Unfortunately, they still cannot create models with correct topology, that can be used in games. But the progress is awesome, especially in Rodin.
Thanks, didn't know the 3D generators had gotten this far. Really need to test it to believe it because I see a base mesh and rough texture. Remesh, sculpt more or new details, weight paint, improve texture and even an inexperienced developer could have a model better than anything I could have dreamt of back in school (many years ago) .
The ability to regenerate a small part of the texture was also really interesting. Would be cool if we could also give a rough model as input to a 3D generator. Either as inspiration or to ask the AI to improve it. Imagine if we could combine mesh painting and a prompt to tell the AI to improve something in a certain way. Ex, give it a WV, highlight the hood on the car and tell it that you want part of the engine to stick out visibly and the AI gives you X number of modified meshes with a visible engine block.
@@KrullMaestaren Rodin has released a feature similar to what you mentioned. I have yet to experiment with it but it looks like you can give some parameters for shaping the mesh. Then combined an image as a reference.
You should make another video where you use the exact same image in each one. It’s hard for a fair comparison otherwise
@@gunterstunter Great idea, I have a couple images cued up from that exact prompt. Definitely would be worth taking a look at
@@ArtificiallyInked glad to hear that! Thanks for the videos, they’re helpful :)
So if i make a reference image for a character with multiple angles, it will work better than multiple images?
@@rwxstudio7173 Yes with tools like Tripo, and Rodin. Multiple images will garner better results. Less guess work the tool has to do.
i can't get cura to open the stl from rodin without upscaling it 10000%. then it won't slice it once resized. all it shows is a shadow. what am i doing wrong?
@@undeadly1103 I have noticed the meshes from Rodin are tiny. I usually scale them up in Blender. I am a noob at 3D Printing so unfortunately don’t have an answer. I have heard there is an extra process to properly convert the mesh so your slicing software will recognize it.
@ArtificiallyInked ok. thanks! i have one from rodin ready to go but I'll put it blender or fusion first and see what happens. 🤞
@ keep us updated. I am curious if these types of models translate well to 3D printing. I am hoping to eventually make some figurines of a few characters on the channel.
I want to be optimistic and say these can be used in a supplemental way. Like if you get concept art for a character with a lot of 2d cheats in its design. The AI can generate model if you get stuck
Exactly, I have worked with a couple teams now that had no art direction. With help from several of these tools and a handful of others they were able to get publishing dollars. Now they have full art teams and projects scheduled for release. These tools are great for bringing the idea to life. Not as much for the product you are trying to deliver to market.
Looks like you can. Treat the result as a base mesh. Remesh the exported result and brake it at the same time into jacket, body, hat, hand etc. I would also do the weight painting on my own to make sure that armor as an example doesn't behave as jelly. You then touch up the textures and apply them to the cleaned up mesh.
You can also add new details to the mesh and texture while working on them.
@@KrullMaestaren You definitely have the concept down for finalizing your characters. This is a great jumping off point but they do need the skilled hand of a 3D modeler to be fully useable. Another advanced trick I have learned is to create head, hands, feet, and body as separate meshes. Rejoin them later in Blender. This will give better results for the mesh and textures versus trying to take one full swing at a whole character.
You could also use a painted image and see how that works. Just a thought. Love the videos though..
@@RyanHarris-ix1lm that is a great idea. Could use an app like Krita that has a Stable Diffusion plugin and draw out a character in real time. That would be a really interesting workflow.
Unfortunately there are no reason to use theese tools on trial - it lacks textures resolution and overall quality, so its difficult to estimate if theese could be helpful in real production, even in the best efforts it lacks complexity, texture accuracy to be compared to human-created stuff, so no reason to pay for tool you cannot use
@@jozinzzzzzbazin I am currently working on update videos for these tools. There has been enough interest and paid subscriptions to improve the output of both meshes and textures for most of these tools. Of course the human element will always be required to fine tune any of the outputs to your standards. Depending on the complexity of your project current outputs can already be put to use in a production.
You used text prompts for these. The title is very deceptive.
@@shaionline th-cam.com/video/8k3QCajRk8o/w-d-xo.htmlsi=60XVTs0lUqTnv7bR
@@shaionline th-cam.com/video/yvNHPCuuzvs/w-d-xo.htmlsi=Y8xfLRcYGHXDTWgm
@@shaionline here are a couple other videos that focus more on IMG - 3D
Whats your favourite nowadays?
@@InfiniteColours with the latest updates. I feel like Rodin is still the best. That tool really dials on the base mesh. Some of the sharpest details yet. It also allows for multiple image references. They just added a face restore feature. I use all three tools for different use cases, but for character work. Rodin wins in its current state
@@ArtificiallyInked thanks! In my experience rodin beats meshy yh, havent tried tripo but seems to be very good. Which one u think is the best for image to 3d? Love ur videos so much, thanks for them!
@@InfiniteColours Tripo is great also. There will be a video out this week showcasing their new update. The new update gives Rodin a run for It’s money with Image to 3D. Where Meshy really shines is Text to 3D. I use all 3 on a daily basis
@@ArtificiallyInked wont miss it for sure haha. Will probly get a sub on both tripo and rodin and test all functions
@ They are definitely useful tools. Especially for developers that maybe wizards at coding but noobs at modeling. Great way to get your projects up and moving
the voice so annoying bro i suggest to find better AI Voice . Thank you for the video bro
@@AIGen-e8k Thanks for the feedback
I like the voice, it fits to the small robot
I honestly liked the voice
nice
Herkese yarratıcılık dılerım