We explain Barkhausen noise in magnetic systems in terms of avalanches near a plain old critical point in the hysteretic zero-temperature random-field Ising model. The avalanche size distribution has a universal scaling function, making non-trivial predictions of the shape of the distribution up to 50\% above the critical point, where two decades of scaling are still observed. We simulate systems with up to \(1000^3\) domains, extract critical exponents in 2, 3, 4, and 5 dimensions, compare with our 2d and \(6-\epsilon\) predictions, and compare to a variety of experimental Barkhausen measurements.