EVR.Systems

EVR x Pixelynx x Deadmau5 We already shared how we developed the graphics for the AR application featuring DJ and producer Deadmau5, but! We didn't reveal details — the features and challenges of the case. 💚 The entire project had to fit within a strict limit of 100,000 polygons. To achieve this, we created custom shaders that simulated geometry and saved resources. Next, we baked texture maps and optimized the ship model in Blender. We scanned 3D objects using LiDAR and created animations in Unity. Additionally, we used a Visual Positioning System to ensure that the interactive content was anchored to the object with an accuracy of up to 10 centimeters. And voila! It turned out to be splendid, dazzling, beautiful, luxurious, awesome, and received an award for the best experiment from Niantic 👀

🤚 Post about our team EVR Systems — connecting brands and their audience using new technologies We create: — Projects in all forms of reality: AR, XR, VR, Meta; Interactive installations of any level: from exhibition stands to stadium shows; — graphics for video: 3D, generative, digital characters; AI, ML, Deepfake and their integration into business processes; — DEV: webGL, mobile, complex websites (3D, video, interactive), integration of pipelines for neural networks, backend development We love to experiment with new technologies, collaborate with other teams, and implement new ideas for business and events 120+ projects for leading brands and global companies 30+ clients: Coachella, Deadmau5, Mahmut Orhan, MTS, Yandex, Sberbank, Pixonic, Gazprombank, Norilsk Nickel, Alfa Insurance, and others 25+ awards: Perspective Awards, Silver Mercury, Effie, Mixx Russia, Red Apple Lightship Summit, Zhara Music Awards, etc 10+ years of expertise and innovation in new technologies We conduct training sessions for teams of managers, creatives, designers, and developers. We share knowledge at conferences. We assemble teams for special projects. We also have a service for renting ready-made installations 💚 Our main feature: a passion for technology and the creation of unparalleled products in the multimedia industry

EVR RnD: SnapFit, Part 2 In the last post, we talked about our development for the fashion industry — virtual fitting rooms. Today, we'll explore the second functionality of SnapFit — AI-generated photos for marketplaces and online stores. What's the catch? The service generates ultra-realistic images of products, saving budgets on photo shoots by a factor of 100 — there’s no need to waste time and money on models, locations, logistics, photographers, lighting, makeup artists, and other now unnecessary expenses. Branding elements — logos, names, and any corporate style — can also be seamlessly and strikingly integrated into the images, making them memorable and pleasing to the eye. About the development: we’re maintaining a good photo reel in the generations. We continue research, refining it into the first stable version, and rolling it out for closed testing. Stay tuned! We 💚 AI technologies! #ai

EVR RnD: SnapFit 💚 We're sharing our internal development for the fashion industry — SnapFit! Internally, we call it the fitting room, and it has 2 cool features — virtual try-on for looks and the generation of selling product cards for marketplaces and online stores. Today, we'll talk about clothing We still have some fine-tuning to do, but what our pipeline produces now looks much cooler than other open-source solutions on the market! It creates a more consistent image — preserving the picture and text from the original clothing better. We pay special attention to photorealism — ensuring that shadows, lighting, folds, and fabric texture work well Of course, it's not zero-shot yet. Sometimes quality generation doesn’t happen on the first try, but on the third, fourth, or fifth — it all depends on the clothing. However, we’re refining the pipeline for stable results and will enhance the technology for consistent outcomes We'll share more about the second function later, but for now, let's admire our tests and comparisons with competitors 👀💚 #ai

EVR Systems. Отраслевые тренды и технологии медиадизайна. У нас в гостях ребята из студии мультимедиа-дизайна EVR Systems Aleksandr Glushchenko co-founder, технический директор EVR.Systems и Anya Krol креативный директор / продюсер. Гости подробно рассказали о своем вдохновляющем пути в сферу мультимедиа, о том, как устроен этот рынок – о типах заказчиков, и среднем чеке за проект. Какой минимальный стек технологий нужен сейчас для входа в рынок, и почему дизайнеру важно разбираться в коде. Ждем вопросы для гостей в комментариях к этому выпуску ;) ССЫЛКИ: 🔗 Ссылка на гайд по тенологиям https://guide.evr.systems/ 🔗 Ссылка на упомянутый кейс https://lnkd.in/eFk2GdFG 👉 Слушать в Apple https://lnkd.in/eaWS8BEP 👉 Слушать в Mave https://lnkd.in/e3qVuD58

Together with @setupdesigns, collaborated with @sensorium_galaxy to develop a show in the metaverse featuring various artists, including @carlcoxofficial Carl Cox's 'Intermundium' is a collaborative exploration where his avatar performs in Sensorium Galaxy 💚🔥 #virtualproduction #virtualreality #xr #oculusquest #quest3 #augmentedreality #AppleVisionPro #applevision #interfacedesign #cyberpunkart #extendedreality #cginteractive #oculusquest3 #visionpro #desksetup #logitech #ar

Exploring the Frontiers of Extended Reality (XR) in Automotive Innovation We’re thrilled to share our latest experiment in the realm of XR, as we redefine the future of automotive presentations and showrooms What sets this XR experience apart from traditional VR is the seamless integration of real-world visuals with 3D effects, offering an exhilarating sense of awe and engagement! #xr #oculusquest #quest3 #augmentedreality #AppleVisionPro #applevision #interfacedesign #cyberpunkart #extendedreality #cginteractive #oculusquest3 #visionpro #desksetup #virtualreality

AI MoCap For one of our projects we train our own AI system with a motion capture. The input consists of one recorded MoCap with skeleton movements and the output includes dozens of generated AI animations, which the neural network combines/creates based on the given prompt You can see our results in the video below

Continuing to share insights into our AI pipeline for texturing assets. Our primary goal was to simplify the workflow—writing prompts can be tedious and challenging. It's much more productive and faster to find an image you like and tell the neural network: "do something similar." One of the texture generation functions involves uploading a reference image and transferring its style to the asset. This approach is swift, allowing us to generate over 500 variations with different images in a day. This provides the flexibility to choose the desired style or explore various concepts. Primarily, we believe this will be beneficial for art directors and designers—an extremely fast way to validate hypotheses and concepts.

How does our generative payplane perform on game assets? Examples on the video, for texturing we need reference images and 3d models.