Visual pigments can be spontaneously activated by internal thermal energy, generating noise that interferes with real-light detection. Recently, we developed a physicochemical theory that successfully predicts the rate of spontaneous activity of representative rod and cone pigments from their peak-absorption wavelength (lambda(max)), with pigments having longer lambda(max) being noisier. Interestingly, cone pigments may generally be similar to 25 fold noisier than rod pigments of the same lambda(max), possibly ascribed to an 'open' chromophore-binding pocket in cone pigments defined by the capability of chromophore-exchange in darkness. Here, we show in mice that the lambda(max)-dependence of pigment noise could be extended even to a mutant pigment, E122Q-rhodopsin. Moreover, although E122Q-rhodopsin shows some cone-pigment-like characteristics, its noise remained quantitatively predictable by the 'non-open' nature of its chromophore-binding pocket as in wild-type rhodopsin. The openness/closedness of the chromophore-binding pocket is potentially a useful indicator of whether a pigment is intended for detecting dim or bright light.