Rendering and Display Algorithms for Large Stereoscopic High Dynamic Range Projections

今天，立体声显示器是一种商品，高动态范围(HDR)显示器已经商业化。这两种特性的结合在技术上也是可行的，例如使用平板显示器和LED背光。然而，当目标是大尺寸显示器(屏幕尺寸在米范围内)时，例如用于可视化或沉浸式应用的大型电源墙，除了受边框影响的平铺屏幕外，没有可行的技术解决方案。我们设想的一种可行的技术实现是将HDR显示系统解耦为立体投影仪(作为源和第一调制器)和反射式电子纸显示器(EPD)作为大投影屏(和第二调制器)。然而，如何在由立体和单视、低刷新率调制器的组合组成的这些设置上显示动态立体HDR内容，这并不是一个清楚而有趣的研究问题。在这个项目中，我们寻求研究能够在大投影屏幕上显示大规模立体HDR内容的算法。原型设置使我们能够研究立体HDR成像和人类感知的各个方面，超越了特定的硬件原型。特别是，我们将探索单视和立体调制器的组合，以改进图像组件的动态范围、异步更新和时间滤波，以及用于交互应用的预测渲染。

Today, stereo displays are a commodity and high-dynamic range (HDR) displays are commercially available. The combination of both of these properties is also technically feasible, e.g. with flat panel displays and LED backlights. However, when targeting large scale displays (screen sizes in the range of meters), such as large powerwalls for visualizations or immersive applications, there exists no viable technical solution except for tiled screens which suffer from bezels. One feasible technical realization that we envision is to decouple the HDR display system into a stereoscopic projector (as source and first modulator) together with a reflective electronic paper display (EPD) as large projection screen (and second modulator). However, it is not clear and an interesting research question how dynamic stereoscopic HDR content can be displayed on these setups consisting of a combination of a stereoscopic and a monoscopic, low refresh rate modulator. In this project, we seek to research algorithms which enable displaying large-scale stereoscopic HDR content on large projection screens. A prototypal setup enables us to research various aspects of stereoscopic HDR imaging and human perception, going beyond the specific hardware prototype. In particular, we will explore the combination of monoscopic and stereoscopic modulators for improving the dynamic range, asynchronous updates and temporal filtering of the image components, as well as predictive rendering for interactive applications.