У нас вы можете посмотреть бесплатно 2025 Advances in RT Rendering - RT Subsurface Scattering via Hybrid RESTIR-Path Tracing & Diffusion или скачать в максимальном доступном качестве, видео которое было загружено на ютуб. Для загрузки выберите вариант из формы ниже:
Если кнопки скачивания не
загрузились
НАЖМИТЕ ЗДЕСЬ или обновите страницу
Если возникают проблемы со скачиванием видео, пожалуйста напишите в поддержку по адресу внизу
страницы.
Спасибо за использование сервиса ClipSaver.ru
https://advances.realtimerendering.co... ABSTRACT: This presentation introduces a novel hybrid solution for real-time subsurface scattering (SSS) that approaches path-traced quality while maintaining interactive performance. Traditional real-time SSS relies on diffusion approximations, but these methods often produce artifacts when handling thin or curved regions, such as nostrils and ears, due to their semi-infinite medium assumption. To address this limitation, we combine a brute-force volumetric path tracing component for single scattering with a newly derived, physically based diffusion profile for multiple scattering. This hybrid approach accurately captures the subtle transmission and scattering behaviors across a wide range of translucent materials, improving visual fidelity and reducing unwanted artifacts. Additionally, we demonstrate how ReSTIR can be integrated into this workflow to accelerate path-traced single-scattering computation. By leveraging ReSTIR’s reservoir-based sampling, our method strategically reuses samples over space and time, reducing noise and computational cost without sacrificing quality. As a result, users can achieve the appearance of path tracing—even in challenging scenarios—at frame rates suitable for real-time applications. Attendees will learn practical strategies for implementing this hybrid approach in their own rendering pipelines. We emphasize how to balance physical rigor with efficient sampling and filtering, ensuring consistent, near-ground-truth results for numerous types of assets and lighting conditions. This talk empowers developers and technical artists to adopt path-tracer-level SSS in real-time environments, marking a significant step toward more lifelike digital humans, film-quality virtual production, and other use cases demanding high-fidelity translucency at interactive speeds. SPEAKER: Tianyi "Tanki" Zhang is a senior real-time rendering engineer at NVIDIA, working on the Omniverse RTX Renderer. As a rendering engineer, his focus includes real-time path tracing light transport algorithms and systems, path tracing real-time neural graphics primitive (Gaussian, NeRF), ultra-realistic XR pipeline integrating path tracing and compositing. Previously, Tanki was a Rendering Engineer Intern at Epic Games, contributing to real-time ray tracing features for Unreal Engine. With a background in computer science, game development, and art design, he is passionate about pushing the boundaries of computer graphics and shaping its future.