Atmospheric oxygen evolution has long been discussed, especially with its relevance to the origin and evolution of life and the planet. Presence/absence of detrital redox-sensitive minerals, iron formations and red beds, behaviors of redox-sensitive elements in paleosols (ancient, subaerially-altered continental rocks) and ratios of carbon and sulfur stable isotopes in sedimentary rocks have been utilized to constrain atmospheric oxygen levels, which can dictate surface redox states, in the distant past, leading to a conventional view that the beginning (2.5–1.8 Ga) and ending (0.8–0.5 Ga) of the Proterozoic were two major periods when the oxygen level significantly increased in the Earth's history. More recent studies adopt multiple sulfur isotopes, iron speciation and trace elements (isotopes) as additional redox proxies. These proxies are not inconsistent with the conventional view, but the magnitude and timing of changes in these proxies are different between proxies and between geological records obtained from, e.g., iron formations, shales and paleosols. Also, the proxies suggest that there may have been transient oxygen increases of uncertain magnitude at 3.3–3.0 and 2.7–2.5 Ga. To better understand atmospheric oxygen evolution, process-based methods which quantify oxygen levels from individual proxies need to be developed to consistently and comprehensively explain multiple geochemical signatures.