This study is intended to measure the geometry of the observer's color space when viewing a computer screen and to define individual variations from this data. A CIE photometric standard observer assumes that the eye's spectral efficiency function is constant, and photometry measurements correspond to vectors with fixed directions. By definition, the standard observer decomposes color space into planar surfaces of constant luminance. Using heterochromatic photometry with a minimum motion stimulus, we systematically measure the direction of luminous vectors for many observers and many color points. During the measurement process, the background and stimulus modulation averages are fixed to the given points in order to ensure that the observer is in a fixed adaptation mode. Our measurements result in a vector field or set of vectors (x, v), where x is the point's color space position, and v is the observer's luminousity vector. In order to estimate surfaces from vector fields, two mathematical hypotheses were used: (1) that surfaces are quadratic or equivalently that the vector field model is affine, and (2) that the metric of surfaces is proportional to a visual origin. Across twenty-four observers, we found that vector fields are convergent and the corresponding surfaces are hyperbolic. The equation of the surface in the display's color space coordinate system, and in particular the axis of symmetry, varied systematically from individual to individual. A hyperbolic geometry is compatible with studies that emphasize a modification of the photometric vector with changing adaptations.