Binocular brightness combinations: Additive and nonadditive aspects

A conjoint measurement procedure is used for the measurement of binocular brightness as a function of left and right luminance inputs. For nonzero stimulation, the data confirm earlier findings: the system can be described as additive with a scale exponent of 1. If zero stimulation is included, however, no additive solution can be found (due to Fechner’s paradox). This fact, combined with various critical remarks in the literature with respect to the existence of a real luminance-averaging system, has led us to propose a model which takes account of Fechner’s paradox, and incorporates “realistic” exponents without requiring a multistage processing mechanism where different levels are characterized by different sensory scales. The proposed model makes the weighting coefficients for the two eyes dependent in a continuous way on the strength of stimulation in the two eyes, especially on the amount of contrast of the monocular stimuli. For zero background stimulation, contrast can be expressed in terms of luminance of the stimulus. In this way, the model is reduced to a simple testable form. While it much simpler than Engel’s (1969) model, the experimental results indicate that it might also work for the more general case.