Calculating meteoroid masses from photometric observations relies on prior knowledge of the luminous efficiency, a parameter that is not well characterized; reported values vary by several orders of magnitude. We present results from an experimental campaign to determine the luminous efficiency as a function of mass, velocity, and composition. Using a linear electrostatic dust accelerator, iron and aluminum microparticles were accelerated to 10+ km/s and ablated, and the light production measured. The luminous efficiency of each event was calculated and functional forms fit for each species. For both materials, the luminous efficiency is lowest at low velocities, rises sharply, then falls as velocity increases. However, the exact shape and magnitude of the curve is not consistent between the materials. The difference between the luminous efficiencies for iron and aluminum, particularly at high velocities, indicates that it is not sufficient to use the same luminous efficiency for all compositions and velocities.