In this paper we present a model for the simulation of the photographic development process for use on computer generated and other digital images.The model provides us with a tone reproduction operator based on photographic principles that mimics the creation process of black and white photographic prints.We focus on four characteristics of photographic materials: density response, spectral sensitivity, resolution and granularity.These characteristics are described quantitatively using empirical data thus making the simulation of the response of actual photographic materials a straight forward application of the model.The result of the simulation is a device independent image of floating point values between 0 and 1 which represe nt shades of gray on a linear scale.This image can be quantized for display on a given output device.

In standard fractal terrain models based on fractional Brownian motion the statistical character of the surface is, by design, the same everywhere. A new approach to the synthesis of fractal terrain height fields is presented which, in contrast to previous techniques, features locally independent control of the frequencies composing the surface, and thus local control of fractal dimension and other statistical characteristics. The new technique, termed noise synthesis, is intermediate in difficulty of implementation, between simple stochastic subdivision and Fourier filtering or generalized stochastic subdivision, and does not suffer the drawbacks of creases or periodicity. Varying the local crossover scale of fractal character or the fractal dimension with altitude or other functions yields more realistic first approximations to eroded landscapes. A simple physical erosion model is then suggested which simulates hydraulic and thermal erosion processes to create gloabl stream/valley networks and talus slopes. Finally, an efficient ray tracing algorithm for general height fields, of which most fractal terrains are a subset, is presented.