We had originally planned or discussed running with R0/ACQ < 1 in autoranging to reduce the contribution of noise in AxL_L from doing CDS subtracting noise/x from noise/L. Should this be revisited? Could future pixels not do CDS in autoranging? What happens in epixM?