The scanning-beam technique for measuring the response of a detector to an
irradiance is analyzed. With this method the irradiance responsivity is det
ermined by integration of the spatial responsivity. Since in practice the i
ntegration is approximated by a summation over steps with a finite step siz
e, errors are introduced. It is shown both theoretically and experimentally
that the error vanishes when the reciprocal step size lies beyond the diff
raction limit. Furthermore, comparison shows that experiment and theory are
in good agreement. (C) 2000 Optical Society of America.