We studied how variation in seafloor water clarity, ambient light, and fish fork length influenced the maximum detection range of fish with a stereo-video lander on three temperate reefs of different depths (12–40, 44–91, and 144–149 m). Although the results are somewhat approximate and specific to the camera system, the methods we used can be applied to any stereo remote underwater visual survey system. In the 52 total lander deployments distributed between nearshore, mid-shelf and deep-shelf reefs in Oregon waters, seafloor light levels varied over 4 orders of magnitude, primarily as a function of depth. The seafloor scattering index was higher (low water clarity) and highly variable at the nearshore reef and lower (high water clarity) and less variable at the deeper reefs. In the 15 deployments with sufficient numbers of fish for detection range analysis, the mean maximum range of detection across species varied from 3.89 to 4.23 m at the deep-shelf reef, 3.32–5.55 m at the mid-shelf reef, and 1.57–3.42 m at the nearshore reef. Multiple regression analysis of the analyzed deployments showed a strong negative relationship between mean maximum detection range and the scattering index but no relationship with log_e of seafloor ambient light. The lack of a light effect showed that the artificial lights were adequately illuminating the field of view in which fish were identifiable, potentially an important system test for sampling across a range of seafloor light levels. Analysis of detection range versus fish fork length for Blue Rockfish Sebastes mystinus and Deacon Rockfish S. diaconus from a single deployment showed a reduction in detection range for 10–20-cm fish of about 1.15 m relative to the detection range of 25–45-cm fish, or about 41%.