Collisions with windows on buildings are a leading direct source of anthropogenic mortality for birds. A common scenario in which birds impact windows is at single-family residences with nearby bird attractants, such as bird feeders and bathing stations. Scientific understanding of bird collisions is limited by a lack of empirical data on how collisions actually happen. Previous studies have documented evidence of collisions after-the-fact (i.e. bird carcasses) but none have observed collision events directly. As a result, many assumptions about bird collisions and methods for prevention have yet to be tested. We developed two methods for documenting, measuring and comparing bird-window collisions in the field and in a laboratory setting. First, we positioned a home security camera system in a residential backyard near bird feeders and passively recorded a variety of wild bird species interacting with windows on the home over two years. From the footage, we estimated birds’ flight speed, angle of approaching the window, and any observable outcome of collision events. Second, we created an indoor flight arena to simulate collisions with windows by having house sparrows (Passer domesticus) impact and bounce off a lightweight transparent plastic sheet. During flight trials, we recorded changes in birds’ flight speed, trajectory and head angle upon approach using high-speed cameras in order to characterize the spatial parameters of birds detecting and avoiding an obstacle. We will discuss findings of analyses of collision footage from the field and captive experiments, and implications for interpreting the causes and effects of collisions with windows at the level of individual birds.