Human contrast sensitivity in coherent Maxwellian view: effect of coherent noise and comparison with speckle

Lasers have been used in vision for measuring the neural contrast sensitivity function (CSF) by forming interference fringes on the retina. We distinguish among three kinds of illumination with lasers: incoherent (without noise), Maxwellian or coherent (with coherent noise), and diffuse coherent (with speckle). The three have different characteristics and different CSF's. A coherent imaging system is designed to measure the CSF with fully coherent illumination. This is the CSF of the whole visual system, although it is measured with gratings imaged on the retina. It therefore differs from the neural CSF's measured by other authors with partially coherent illumination. However, the neural CSF's are also obtained in this study with and without noise. The effects of coherent noise and speckle on both the visual system and neural sensitivities are studied and compared. Coherent noise differs from speckle in the following ways: (1) It behaves as a high-pass filter, reducing sensitivity in the low-spatial-frequency range, whereas speckle is a low-pass filter; (2) quantitatively, coherent noise reduces neural sensitivity by a factor k(m) with a maximum value between 4 and 6, whereas speckle reduces neural sensitivity by a factor ks with a maximum value of approximately 25 (1.4 log units) for a 3-mm pupil and up to 35 (1.55 log units) for a 1-mm pupil; (3) the masking effect of the coherent noise is affected by changes in luminance but not by changes in pupil diameter; however, the pupil size is the main parameter affecting the masking effect of the speckle.