This paper deals with measurement and characterization of IR sensor performance. Alternative methods have been investigated in order to determine the performance more reproducibly and accurately. Furthermore, the new methodologies reduce the subjectivity of the performance parameters such as the standard minimum resolvable temperature difference (MRTD) and tend to reduce the time (and thus cost) of measurement. The three methods described concern the use of (1) and IR zoom collimator, (2) a test device with line targets, and (3) an extended target combined with a TV camera instead of a human observer and a double-line source, resolving the ambiguity beyond the Ny-quist frequency in staring focal plane array cameras. In the IR zoom collimator a variety of targets are used in four different orientations; the observer has to report not only the recognition, but also the orientation of the target. Instead of variation of the temperature difference of the target, its range is varied by zooming in and out. In the rapid objective MRTD tester, the performance of the eye is simulated with a simple formula. The complete MRTD is obtained in about two minutes. In the double-slit method the best phase of the detector matrix is taken in comparison with the two line sources; instead of the MRTD curve, a new type of performance curve is measured: The minimum resolvable intensity difference (MRID), from which the range performance is obtained similarly to the classical MRTD method. The paper contains examples of measurements with each of the methods. ©2003 Society of Photo-Optical Instrumentation Engineers.