Abstract

The resolution of 3-bar targets is an established criterion for the quality of optical systems; however, most optical design programs calculate the OTF function rather than the 3-bar resolution. In order to be able to predict the resolution of a system before it is built and experimentally tested, a conversion of the OTF data to resolution data would be useful. This paper deals with the problems associated with this conversion. A calculation method is given which correctly accounts for the following properties of 3-bar imaging: square-wave vs. sine-wave object; finite vs. infinite number of cycles in the pattern; linear systems theory vs. non-linear film behavior.

It is shown that the use of a threshold function for the MTF can lead to incorrect results, since the threshold function method does not take the shape of the MTF curve into account. Furthermore, it is shown that when 3-bar resolution is measured using the ANSI standard resolution chart, the resolution of a particular pattern may be significantly affected by the presence of neighboring patterns. It is seen that the use of highly non-linear film for resolution testing can lead to inaccurate measurements.

© 1994 Optical Society of America