Optical system designs specifically tailored for the high apertures required for direct view, passive, image intensifier, night vision systems began in the late fifties with a 14 inch focal length F/1.4 catadioptric objective. Three spherical corrector lenses were used in a flat field design for a 40mm image intensifier system. The principles developed proved quite successful, eventually leading to similar objectives used on most 40mm, 25mm, and 18mm format, first generation, night vision sighting telescopes. The principles are discussed and comparisons are made with several other high aperture catadioptric systems recently discussed in the literature. More compact second generation electrostatic and wafer type tubes opened new opportunities for light weight system designs. The 18mm microchannel plate wafer tube in particular resulted in a number of new viewing systems, including nairs of 1X wide angle telescopes used as night vision goggles. The nrinciples involved in these systems are described as well as the eyepiece performance and characteristics which are unusually important. For these eyepieces and in fact for all of the newer Magnifier eyepieces, Night Vision Laboratories has developed new methods of testing and criteria for specifications that require considerably better correction that had previously been required. The most recent development in the field of direct view night vision devices has been that of the Biocular eyepiece. The principles and characteristics of this rather successful innovation are discussed.

The observation of objects under low illumination conditions is made possible using photo-electronic image intensification of the available light. The intensifier is normally used in conjunction with a large aperture objective lens so as to gather the maximum number of photons. These photons are then imaged onto the photocathode and after electronic amplification the reconstituted optical image can be viewed at the phosphor by means of an eyepiece. At present the limitation of passive night vision devices is the inadequate spatial frequency response of the image intensifier.

The multi-mirror telescope project is a joint undertaking of the University of Arizona and the Smithsonian Astrophysical Observatory. The telescope, to be located on Mt. Hopkins south of Tucson, Arizona, will be used primarily in the fields of infrared photometry and spectroscopy. Although methods for tolerancing optical systems, especially astronomical telescopes, exist, salient features of the Multi Mirror Telescope (NMT) force us to use a new approach to the tolerafting. Before we discuss these methods, a general description of the MMT is in order.

The manufacture and testing of a silver parabolic mirror cut directly with a diamond tool in a high precision numericallycontrolled lathe without subsequent optical figuring or polishing is described. The mirror is 6 in. in diameter with a 1.5 in. facal length, giving an f-number (ratio of focal length to beam diameter) of 0.25.

This paper is a summary of some remarks made at the 1973 Seminar on Geometrical Optics, regarding the state-of-the-art and the benefits and problems associated with the interferometric testing of optical systems.

"Real time" is an expression commonly used to mean that we learn what is happening while it's happening. By modifying that expression, the topic of this paper might well be "real life" testing of optical components, or "learning whatever happened to that piece of glass".

Obviously, if the eye forms part of an optical system, the properties of the eye must be taken into account in assessing the performance of the system, and even in designing it in the first place. The unaided eye is far from being diffraction limited, and it has many other properties which affect the design of any optical system in which the eye is the ultimate image receiver. We shall consider here three particular aspects of the human eye: its di-mensional properties, its photometric properties, and its resolving power.

The task of designing an ophthalmic corrective lens is one of optimizing a system by modifying only a fixed front element with remainder of the system rotatable about the stop point. Commercial designs date from the first part of the century. Two schools of thought, which still persist, governed the thinking and design work: Whether to correct astigmatism or to control field curvature. These meridional off axis errors have been the primary concern of ophthalmic lens designers, and considered mainly for distant object points. Geometric Optics has played a minor role in bifocal design except for research in the area of unsegmented multifocal lenses of continuously varying power. A design project today, made practical by computers and plotters, is one of optimizing a series of several hundred compatible lenses with consistent, logical tolerances, trade-offs and priorities applied to a range of fitting and object dis-tances. The magnitude of errors involved varies with the task from a few hundredths of a diopter in some prescriptions to several diopters in aphakic and low vision corrections. Aspherics work wonders with such lenses. The acceptance of hard resin lenses, has also made aspherics practical for ordinary prescriptions. This allows consideration of additional criteria such as distortion and cosmetic appearance.

With "ordinary" optics,i.e., refractive and reflective optics, obtaining field of view is not a problem, per se. Holographic optics, however, are diffractive in nature; in order to achieve the required optical efficiency, one usually must work with "thick" or "volume" holograms, and these volume holograms have a strong angular selectivity that tends to limit the field of view.

Conventional optics normally includes only two types of elements: refractive and reflective. Holographic optics encompasses a third type of element; one whose properties are diffractive in nature. The focus of this paper is to outline the properties of a general computer-based analysis and design tool that is directed to holographic optics. This tool is called the Holographic Optics-Analysis and Design (HOAD) program. The program has been designed to be both user oriented and flexible. Four sections are covered in the paper. They are directed to-wards (1) the operating system (MTS) used to support the HOAD program, (2) the hologram ray tracing programs which form the basis of the program, (3) a general discussion of the HOAD program, and (4) an outline of the techniques of using HOAD. The last section of the paper is directed towards an assessment of where HOAD presently stands and the direction it is going.

It is well known that the first-order properties of an optical system are sufficiently specified by the tracing of two meridional paraxial rays through the system, namely the marginal ray, which originates at the center of the object plane and passes through the edge of the stop, and the chief ray, which originates at the edge of the object and passes through the center of the stop. Wherever the marginal ray crosses the axis an image plane is located and wherever the chief ray crosses the axis a pupil plane is located. Moreover, the height of the chief ray at an image plane gives the size of the image and the height of the marginal ray at a pupil plane gives the size of the pupil.

Matrix optics is currently enjoying great popularity as a pedagogical tool, but the technique has not been widely accepted by optical designers, probably because it offers no advantage for numerical computation over ynv ray tracing. The essence of matrix optics, however, is not that it provides a tool for numerical computation, but that it provides an abstract formulation of first-order optics which is at the same time easy to work with and penetrating in its analysis of paraxial optics as a linear system. Numerical values of the matrix elements, if required, can be determined by ynv ray tracing. Such a formulation leads naturally to the introduction of abstract ray coordinates analogous to Buchdahl's paracanonical coordinates, s and t. This, however, often leads to theories which are elegant, but which are very difficult to understand. Our present study suggests that a useful compromise between this, and the straightforward method described in elementary textbooks, can be achieved by a judicious choice of reference planes used for the definition of the system transfer matrix. In particular, if the transfer matrix is taken between the entrance pupil and the last refracting surface, then the matrix elements yield the focal length, the back focus, the pupil mag-nification, and the exit pupil location directly.

An analytical description of the y-S7 diagram is given in terms of plane vectors representing the points and lines in the diagram. This allows for a very powerful and elegant tool for deriving and using the design and synthesis properties of the y-37 diagram. Exact analytical expressions are given to derive the first-order properties of an optical system from its representation in the diagram. The Seidel (third order) aberation coefficients are expressed in terms of variables better suited to the y-Yr diagram, and the description of a computer program that implements these ideas, is included. At the request of a number of participants in the Seminar, a complete bibliography on the y-y diagram is included as an Appendix.

The eccentric pupil ("off axis") paraboloid is the most pop, ar optical system for the infrared because there are no obstructions in the light path. However, the field of view is limited owing to coma. Better overall image quality can sometimes be obtained by using a spherical mirror such as found in the quasi-Schmidt systems. Image quality not withstanding, we often find that we need a long focal length in a short space, and so we look toward more elaborate optics than a single mirror. Two-mirror reflectors, such as the Cassegrain, can be made with an eccentric pupil so that there is no obstruction, or with a Gregorian (concave secondary, intermediate image) reflector, which is commonly used in infrared telephoto systems.

Part of the work done under an Air Force contract, and previously under Lockheed's internal funds, has been to find an optical system covering ten's of degrees with arc-seconds resolution. Since we only required a one-dimensional field-of-view, we allowed nonrotational symmetric figuring! Thus we could design an f/2 refractive-corrector-Schmidt-type system that would meet re-quirements. It is a factor of 10 better than a standard Schmidt. The same principle allows a reflective-corrector-Schmidt with better performance than a standard Schmidt. We have analyzed a variety of multiple corrector configurations with outstanding performance, including an f/2 all-reflective system that fully meets requirements, and an f/l. 3 three-refractive-corrector system that exceeds resolution requirements by almost an order of magnitude.

Many of us tend to take for granted the usual optical dictum that non-symmetrical optics cannot be made in a routine way. While, literally, this may be true, non-symmetrical, non-spherical optics can be produced in quantity in a controlled manner; we are doing so at Tinsley today.

Problems associated with scientific uses of high-speed photography are many. This paper deals with the problems encountered in using long focal length relatively fast objective lenses at short conjugates (less than 20:1) and gives one solution. The problems arise when small subjects must be photographed from many feet away. A long focal length lens (500 MM or longer) is required in order to achieve enough magnification to resolve fine detail (thousandths of an inch) and a reasonably fast speed (f/5.6 or faster) is required when a non-light emitting subject is photographed at very short exposure time (millisecond or microsecond range). All lens aberrations are corrected for infinity unless otherwise stated by the manufacturer. At conjugates of approximately 12:1 aberrations start to degrade the image noticeably when fast relative apertures are used. An ellipsoidal mirror corrected for the conjugates and f number needed, is one solution. Subjects covered are: determining when an ellipsoid is required, how it should be specified, how it can be tested, and design and performance of a completed system.

Rotational symmetry in lenses is such a powerfully simplifying property that the consideration of unsymmetric systems for straight imaging applications has rarely been considered. In the past, this natural prejudice has been bolstered by the belief that unsymmetric systems would be impossible to design, or at least impossibly difficult, and even if a design were possible, it probably could not be made. Recent technological advances, such as the computer and the laser, have somewhat weakened such arguments as a rationale for ignoring this problem, however, and some investigations into it are beginning to be carried out.

Our purpose is to present and resolve as best we can various prob-lems that we have encountered in design of zoom lenses by computer optimization. During the past several years we have tried to develop for our own purposes some basic understanding of zoom lenses comparable to the knowledge general to lens designers of how a fixed focus lens works.

There are many lenses, particularly those used in photographic applications, which are focused for different object positions, forming images at different magnifications. Moving the entire lens assembly along the lens axis while the object or image locations are unchanged is termed a conjugate shift, and because the lens aberration correction is well maintained for a small conjugate shift for most lenses, this has become the traditional focusing technique. Newly designed, compact telephoto and wide angle lenses, however, do not have a stability of correction over a sufficiently large range of conjugate distances to satisfy the needs of the photographer, and thus employ various mechanical techniques to compensate for the aberration change. In addition to these "extreme" lenses, the question of how much normal lenses degrade in performance over a large conjugate change, say m'.0 to m'.-.3 arises, followed by the question of whether normal lenses can be designed which will be uniformly corrected over such large ranges. This paper discusses the limitations of some popular lens types, and shows some methods which can be employed to achieve the desired uniformity.

The principal function of the lens designer is the creation of an optical system which yields stigmatic imagery over a specified field of view, and often, over a wide spectral range. The modern lens optimization programs usually permit the designer to refine the mono-chromatic correction to a surprising degree, and choice of the optimum glass types will, without a doubt, result in further improved correction. Superior chromatic correction will probably never obtain, however, if glass types are not selected judiciously At the outset of the design effort. This is true of both field-dependent and axial types of chromatic defects, but the field-dependent defects are often more amenable to reduction through minor adjustments in system geometry. Reduction of axial chro-matic defects usually requires that major alterations be made in the configuration and this discussion will be restricted to reduction of only these image errors.

A concave hemispherical mirror with a segment of a smaller concentric spherical mirror at the paraxial focal region of the larger mirror can be used as a retroreflector having some advantages over the commonly used corner reflector, i.e. wider angular response and lighter weight, the latter because no glass or other dielectric is required. Such a device has been successfully designed and built for microwave applications,4 and may also have utility for infra-red or optical systems requiring echo�enhancement. To determine the optimum radius of the smaller reflector relative to the larger hemisphere, a geometric optics analysis was performed with the aid of a UNIVAC 1107 computer. This gave a better understanding of the effects of spherical aberration on the echoing performance of the reflector system and helped to determine the size of the smaller reflector which gives the best compromise between reflectivity and aperture blockage.

The design of hand-held stabilized sighting devices has generally involved the addition of elements to the optical system, which in turn resulted in the sacrifice of some aspect of performance. Such sights have also tended to be large, heavy and require power. Previously little effort has been directed towards achieving stabilization with the existing elements of the monocular. This has been accomplished in the system described here, with no optical compromises, by using the existing erecting prisms as the stabilizing elements. In addition the system is passive; i.e. it requires no power for operation. The result is a low cost, stabilized, high magnification monocular for the consumer market; a combination that had not been achieved before.

A 0.5 m Ebert spectrometer of high sensitivity (Figure 1) was de-signed, fabricated and calibrated by the Johns Hopkins University. It was mounted on Apollo 17 for research on the ultra-violet spectrum of the lunar atmosphere, with W. G. Fastie of the Physics Department as Principal Investigator. The Ultra-Violet Spectrometer (UVS) 5169 was designed for operation in the range 1175 A to 1675 X, but had zero transmission for all wavelengths in air. A method was developed for aligning subassemblies (using visible light) in such a way that the assembled instrument would be aligned in the ultra-violet to an accuracy of 2 with maximum efficiency, In the Vacuum Optical Bench (VOB), the efficiency and wavelength calibration were determined. If required, a small AX correction could then be made with a precision of 0.5 which would be verified during final VOB calibration. An alignment procedure using precision optical tooling will be described. Enough of the optical and mechanical features will be given to show how the basic objective was attained, and also how the inherent flexibility of the instrumentation proved useful in the rapid diagnosis and correction of several problems which were encountered. This program was a joint effort by the Physics Department and the Applied Physics Laboratory of the University. The UVS was mounted in the SIM bay of Apollo 17 (Figure 2), where it could be operated in several modes for specific investigations. These include observing:1. resonance re-radiation from the lunar atmosphere below the command module against the dark lunar surface just beyond the sunset line; or above the orbit by rolling the module 180°; 2. the UV reflectivity of the lunar surface under solar illumina-tion; 3. galactic Lyman Alpha (Lc)radiation scattered from the moon's surface on the night side; 4. galactic La during translunar coast; 5. the hydrogen halo around the earth during trans-lunar coast. The UVS S169 operated successfully during December 1972, with results to be published by the Principal Investigator.

The title of my talk was chosen because in 1966 the Institute of Optics had a Conference on Lens Design With Large Computers. This conference was arranged by Dr. W. L. Hyde and it turns out that it now appears that it was at the peak of the excitement involved in learning how to harness the modern large computer for lens design. Since that heyday, lens designers and interest in lens design has certainly been muted. This present conference is a timely reminder that this remarkable field of physics and engineering is strong and healthy. The people in the field can weather severe economic problems and the uses of lenses in modern technology is steadily in-creasing.

This survey of image evaluation by numerical techniques makes no pre-tense of surveying the entire field of image evaluation. It has been directed to some of the subjects considered to be particularly timely and of broad interest. The numerical techniques involved in the solution of various problems in image evaluation or image analysis will be covered.

In computer assisted optical design, it is necessary to have a single scalar function which represents the overall quality of the lens, a quantity which may be used as the basis for comparison by the program. Although this is normally referred to as the "merit" function, it is usually based on the defects in the image. Hence, the best lens has the smallest "merit", and the optimization procedures seek to minimize, the value of the merit function.

Different points of view which have been used by lens designers who have written automatic lens design programs are reviewed with emphasis on the idea that some points of view are relatively more productive of useful concepts for the lens optimization problem. Simplification of the linearized problem by linear transformations of the construction variable set and the defect set is discussed. With simplified linear equations some aspects of the nonlinear problem become more obvious, in particular, a separation of the problem into two parts is revealed which is important to the develop-ment of efficient algorithms.

In order to minimize the overall length and diameter of a zoom objective system with non-trivial magnification change, when the entrance pupil must be located in front of the first element, an intermediate image should be formed, a field lens located at or near that image, and a zoom relay used to produce the focal length change. Usually such lens systems compensate image shift by moving all or part of the front element or by dividing the zoom relay into two independently moving parts. In this paper we describe an improved approach in which the tasks of focal length change and image shift compensation are shared by movements of the field lens and of the relay. This facilitates the optical design, allows the internal pupil to be favorably located, creates the possibility of fixing the exit pupil in size and position, reduces the overall size of the system, and allows a more favorable optical and mechanical design of the compensating parts. The advantages of the new concept are discussed, thin lens examples given, and potential applications to specific types of instruments suggested.