Frame Manipulation Techniques in Object-Based Rendering
This paper is an analysis of rendering methodology based on high-level abstractions. We make a step further from the standard graphics pipeline architecture and include logical objects as a new primitive type into rendering process. The aim of the propose
- PDF / 888,986 Bytes
- 9 Pages / 439.37 x 666.14 pts Page_size
- 105 Downloads / 231 Views
)
Belarusian State University, Minsk, Belarus [email protected], [email protected]
Abstract. This paper is an analysis of rendering methodology based on highlevel abstractions. We make a step further from the standard graphics pipeline architecture and include logical objects as a new primitive type into rendering process. The aim of the proposed method is to simplify rendering process defi‐ nition by reducing the amount of code required to express programming intents. In particular we focus on problems related to frame manipulations and demon‐ strate the advantages of high-level abstractions over pure pipeline implementa‐ tion. Keywords: Graphics pipeline · Rendering · Object-based rendering
1
Introduction
Computer graphics is a very rich technological area, which sees constant improvement on techniques and tools, as well as increasing number of applications. Graphics hardware essentially stays on the cutting edge of parallel computing technologies and rapidly expands its use even to areas unrelated to visualization. However the situation with graphics development complexity does not improve, if we consider standard graphics API, such as DirectX and OpenGL. The level of abstraction where API cuts off stays very low, for better or worse. On the one hand, the interface remains maximally flexible so it can be applied very broadly, on the other hand, it can be considered too low for any meaningful direct application. This particularity of graphics API has resulted in creation of a vast variety of graphics engines [1] – specialized high-level graphics frameworks providing more convenient visualization abstractions, ready for immediate use. Graphics engines are great tools for graphics development, however they trade pipe‐ line’s flexibility for convenience, and implement often a rigid model of visualization process that can fall short in terms of performance or reasonable applicability under particular circumstances. The most technically advanced engines also remain commer‐ cial secrets of corresponding companies. This demonstrates that graphics pipeline and graphics API alone are not enough for high performance and quality visualization, and the accumulated experience could be used to improve the underlying architecture.
© Springer International Publishing AG 2017 V.V. Krasnoproshin and S.V. Ablameyko (Eds.): PRIP 2016, CCIS 673, pp. 97–105, 2017. DOI: 10.1007/978-3-319-54220-1_10
98
V. Krasnoproshin and D. Mazouka
1.1 Related Works Improvement of graphics engines architecture has always been in the focus of computer graphics community. Developers from nVidia [2] and Dice [3] recognize the necessity of changing the common scene graph model into something oriented more to the processes of data transformation. This idea primarily comes from the perspective of parallel rendering, which is essentially a cornerstone of graphics hardware, but not so much of a software. A promising methodology of efficient data-oriented rendering is described by Haaser et al. in “An Incremental Rendering VM” [4]. They show exactly how hig
Data Loading...
