Today, I was working on a prototype when I needed a chest asset.
Instead of doing like I always do (blender or Asset Store), I decided to try out the new AI tools for 3D content generation. This video is a step-by-step of my experiment.
-----------------------------------------------
10 Steps:
1. Go to Sloyd.AI : app.sloyd.ai/
2. Create the asset you want.
3. Make some tweaks to the model, or not.
4. Export the model.
5. Go to TEXTure and upload the model : texturepaper.github.io/TEXTurePaper/
6. Write a prompt for your texture
7. Click Run
8. Download the new texturized model.
9. Import it in your favorite game engine
10. Use it in your game!
-----------------------------------------------
To accelerate the video-making process, I used other AI tools:
I cloned my voice and then used Text-to-Speech using ElevenLabs: lnkd.in/eEdRNCxv.
I used stable diffusion for the coffee image and the thumbnail image: lnkd.in/eqE46sc2.
For the background music, I used Aiva.Ai: creators.aiva.ai/. Credit to them.
Sloyd.ai does not support text to 3D yet, but they will soon. Instead, you could generate the 3D mesh with Shap-E: lnkd.in/e_SHJRxF.
Conclusion:
It is still not as fast and qualitative as the current options to make 3D assets. However, it is already impressive, and maybe in a few months, we will get better and faster results! Pretty interesting.
To all fellow game developers, I hope these AI tools can help us achieve more!
コメント