[{"content":"Introduction All of us love to play video games. From massive titles like Black Myth: Wukong to indie favorites, game studios spend a significant amount of time and money, designing and creating 3D assets. With the growing adoption of VR and AR, the demand for 3D assets is only going to increase.\nAccelerating 3D Asset Creation Currently, we have a few techniques to accelerate the process, but not all of them are widely used yet.\nText-to-3D and Image-to-3D We have AI tools and models like Meshy.ai, SAM-3D, and others that take text or images as input and generate 3D models.\nProblems:\nText-to-3D: It is very difficult to properly describe a complex 3D model using text alone. Image-to-3D: While image inputs offer better control allowing us to generate or edit images with AI tools—most tools rely on a single image. If you upload a front view of an object, the model has to guess what the back looks like, and hence lesser control. Multi-view image-to-3D tools are also available but we haven\u0026rsquo;t got a chance to experiment with it yet. Mesh Quality: A major common problem with these tools is that the generated meshes lack clean geometry and topology, and often have excessively high polygon counts (dense irregular vertices, loose points, and unoptimized geometry). Even if it looks good, it becomes difficult when they want to do edits to it. Photogrammetry It is a technique widely used in AAA game production (your favorite game likely used it). It creates realistic 3D assets by taking multiple photos of an object from different angles and using software to stitch them together into a 3D model.\nThis technique is so effective that in the video below you can see AEOS Games booked an entire town to capture photogrammetry scans, clean them up, and use them in their game. They explained that photogrammetry is the main reason their small team was able to build a really good game (Unleash the Avatar).\nPhotogrammetry discussion starting at 11:47 The way it works is that you take multiple photos of an object from different angles. The software tracks key features across images, generates a point cloud, and reconstructs a 3D mesh from that point cloud.\nLimitations: Photogrammetry struggles with reflective surfaces and requires stable, consistent lighting. Overall, it\u0026rsquo;s a pretty useful technique.\nVideo-to-3D Photogrammetry is great for real-world objects, but if we want to create custom sci-fi assets that don\u0026rsquo;t exist in real life, it won\u0026rsquo;t work.\nThis was true until recently. But what if we used AI video generation models to bridge the gap? So we went ahead and tried it, especially since photogrammetry is already integrated into studio pipelines, this could offer a viable workflow for generating quick base meshes.\nWe decided to test this workflow, and here\u0026rsquo;s what we did:\nConcept Image: Used Nano Banana Lite to generate an image of Pikachu. Synthetic Video: Used Omni Flash to generate an orbital 360-degree video from the image. Reconstruction: Used online photogrammetry software to convert the generated video into a 3D model. Below are the generated video, and the resulting 3D model:\n3D Model • Drag to rotate Interactive 3D Pikachu base mesh generated from synthetic AI video (drag to rotate) While the model needs some cleanup and topology fixes, it looks like a pretty usable base mesh. We also tried it for a Sci-Fi Helmet and for a character from Dragon Ball Z and the results are below\nCurrent Major Limitation: AI video generation models cannot generate long, perfectly consistent videos yet, making full 360-degree coverage difficult for complex objects, larger assets. However, with the rapid development of world models (like Google DeepMind\u0026rsquo;s Genie), generating longer and more spatially consistent videos might quickly become a reality.\nWe also experimented with Gaussian Splatting for a scene, but due to the short duration of the AI-generated video clips we were not able to get the model to generate a proper video, so the results were not good.\nDrag to rotate Interactive 3D Gaussian Splat scene generated from synthetic AI video (drag to rotate) A Benchmark for Video Generation Models ? The more spatially consistent and detailed an AI video is, the better the final 3D model will be. Because of this, converting generated videos into 3D meshes could serve as a one of many practical benchmark for evaluating the temporal and geometric consistency of video generation models.\nConclusion We aren\u0026rsquo;t 3D technical artists, but we do think this is a cool fun idea to try out and hence we tried. We would love for people to try it out themselves and see it might be helpful for them.\nDisclaimer: Pikachu and all associated character names, logos, and designs are trademarks and copyrights of Nintendo, Creatures Inc., and GAME FREAK. All references and generated assets in this article are used strictly for non-commercial, educational, and experimental demonstration purposes.\n","permalink":"/posts/are-ai-videos-the-right-direction-for-3d-asset-generation/","summary":"Exploring the viability of AI-generated videos for 3D asset generation.","title":"Are AI Videos the Right Direction for 3D Asset Generation?"},{"content":"Who We Are We are just a bunch of engineers and creators who love building projects, experimenting with new technologies, and having fun doing it.\nWhy byroot? We decided to put our work and thoughts out into the world hoping to connect and collaborate with like-minded people.\nWhether it\u0026rsquo;s open-source tools, technical deep dives, or experimental side projects, this space is where we share our journey, learnings, and ideas.\nLet\u0026rsquo;s Build \u0026amp; Collaborate You don\u0026rsquo;t need a budget, a corporate setup, or a formal pitch to reach out to us. We build because we love technology, and we\u0026rsquo;re excited to work on cool, meaningful, or fun software projects completely free of any commercial pressure.\nWe love to help out, share advice, and collaborate whenever we can (as time and bandwidth allow!):\n🚀 Building Custom Projects \u0026amp; AI: Have an idea for a cool app, AI workflow, LLM integration, or full-stack project? We love taking on interesting builds and bringing ideas to life alongside you whenever we have the bandwidth. 🤝 Co-Building \u0026amp; Collaborations: Looking for a co-builder or contributor for an open-source tool, side project, or experiment? We\u0026rsquo;re open to teaming up whenever possible. 💬 Help, Advice \u0026amp; Guidance: Stuck on a stubborn bug, architecture choice, or just need a second pair of eyes? Ping us anytime—we\u0026rsquo;re happy to jump in and help out whenever we can! Get in Touch \u0026amp; Socials Have a project idea, need help, or just want to say hi? Reach out across any of our channels:\nEmail: reach@byroot.dev Github: byroot-dev X (Twitter): @byroot_dev LinkedIn: byroot-dev ","permalink":"/about/","summary":"\u003ch2 id=\"who-we-are\"\u003eWho We Are\u003c/h2\u003e\n\u003cp\u003eWe are just a bunch of engineers and creators who love building projects, experimenting with new technologies, and having fun doing it.\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"why-byroot\"\u003eWhy byroot?\u003c/h2\u003e\n\u003cp\u003eWe decided to put our work and thoughts out into the world hoping to connect and collaborate with like-minded people.\u003c/p\u003e\n\u003cp\u003eWhether it\u0026rsquo;s open-source tools, technical deep dives, or experimental side projects, this space is where we share our journey, learnings, and ideas.\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"lets-build--collaborate\"\u003eLet\u0026rsquo;s Build \u0026amp; Collaborate\u003c/h2\u003e\n\u003cp\u003eYou don\u0026rsquo;t need a budget, a corporate setup, or a formal pitch to reach out to us. We build because we love technology, and we\u0026rsquo;re excited to work on cool, meaningful, or fun software projects completely free of any commercial pressure.\u003c/p\u003e","title":"About"}]