Real-Time BTF Rendering

In proceedings of The 8th Central European Seminar on Computer Graphics, pages 79-86, Apr. 2004
Presented at The 8th Central European Seminar on Computer Graphics
 

Abstract

The Bidirectional Texture Function (BTF) is a suitable representation for the appearance of highly detailed surface structures under varying illumination and viewing conditions. Since real-time rendering of the full BTF data via linear interpolation is not feasible on current graphics hardware due to its huge size, suitable approximations are required in order to reduce the amount of data and to exploit the features of current graphics hardware.

In this paper we exploit the recently proposed Local Principal Component Analysis (LPCA) compression which provides a reasonable trade-off between visual quality, rendering speed and size of the representation and present a new rendering method which is approximately three times faster than the originally proposed method. This is achieved by reparameterizing and rearranging the BTF-data appropriately so that built-in hardware interpolation between the values of the discrete LPCA representation can be exploited.

Keywords: color, shading, shadowing, texture program

Images

Download Paper

Download Paper

Bibtex

@INPROCEEDINGS{schneider-2004-real-time,
      author = {Schneider, Martin},
       pages = {79--86},
       title = {Real-Time BTF Rendering},
   booktitle = {The 8th Central European Seminar on Computer Graphics},
        year = {2004},
       month = apr,
    keywords = {color, shading, shadowing, texture program},
    abstract = {The Bidirectional Texture Function (BTF) is a suitable representation for the appearance of highly
                detailed surface structures under varying illumination and viewing conditions. Since real-time
                rendering of the full BTF data via linear interpolation is not feasible on current graphics hardware
                due to its huge size, suitable approximations are required in order to reduce the amount of data and
                to exploit the features of current graphics hardware.
                
                In this paper we exploit the recently proposed Local Principal Component Analysis (LPCA) compression
                which provides a reasonable trade-off between visual quality,
                rendering speed and size of the representation and present a new rendering method which is
                approximately three times faster than the originally proposed method. This is achieved by
                reparameterizing and rearranging the BTF-data appropriately so that built-in hardware interpolation
                between the values of the discrete LPCA representation can be exploited.},
  conference = {The 8th Central European Seminar on Computer Graphics}
}