4 Radiometry, Spectra, and Color

To precisely describe how light is represented and sampled to compute images, we must first establish some background in radiometry—the study of the propagation of electromagnetic radiation in an environment. In this chapter, we will first introduce four key quantities that describe electromagnetic radiation: flux, intensity, irradiance, and radiance.

These radiometric quantities generally vary as a function of wavelength. The variation of each is described by its spectral distribution—a distribution function that gives the amount of light at each wavelength. (We will interchangeably use spectrum to describe spectral distributions, and spectra for a plurality of them.) Of particular interest in rendering are the wavelengths ( lamda ) of electromagnetic radiation between approximately 380 nm and 780 nm, which account for light visible to humans. A variety of classes that are used to represent spectral distributions in pbrt are defined in Section 4.5.

While spectral distributions are a purely physical concept, color is related to how humans perceive spectra. The lower wavelengths of light ( lamda almost-equals 400 normal n normal m ) are said to be bluish colors, the middle wavelengths ( lamda almost-equals 550 normal n normal m ) greens, and the upper wavelengths ( lamda almost-equals 650 normal n normal m ) reds. It is important to have accurate models of color for two reasons: first, display devices like monitors expect colors rather than spectra to describe pixel values, so accurately converting spectra to appropriate colors is important for displaying rendered images. Second, emission and reflection properties of objects in scenes are often specified using colors; these colors must be converted into spectra for use in rendering. Section 4.6, at the end of this chapter, describes the properties of color in more detail and includes implementations of pbrt’s color-related functionality.