The Super Touch 3-D/Optimax III system went through several iterations during its decade of active use. The exact construction of Michael Findlay’s earliest unit is not precisely known, but later designs by Bill Bukowski and by Carl F. Fazekas for Panavision give a clear picture of the basic principles.

Findlay and his successors faced the challenge of getting two Techniscope-format frames shot through two lenses with optical axes about 6.5cm (2.56 in) apart onto one band of film in an over-and-under, or “stacked”, configuration. Offsetting the optical axes horizontally was simple, but offsetting them vertically presented a challenge. Additional mirrors or prisms in the light path could diminish image sharpness, brightness and contrast – in addition, some fairly obvious schemes were excluded by existing patents. Findlay decided to sidestep the issue altogether by making one brazen compromise: the right-eye prime lens in his Super Touch 3-D unit would be slightly higher than the left-eye prime. While the center-to-center distance of the images as measured on film amounts to less than 1cm (0.394 in), Bill Bukowski’s patent allows that the lenses could be as much as 13–14mm (0.512–0.551 in) off level (Bukowski, 1984).

This vertical offset in one lens introduces a condition known as vertical parallax. (Note that this is distinct from “vertical misalignment”, which typically involves a uniform vertical offset across an entire image field, unaffected by camera-subject distance.) The left and right images become increasingly out of vertical register as subjects get very close to the camera, but also as subjects move away towards “infinity”. Vertical parallax is milder for subjects in the middle ground, and not present at all at the point where the optical axes of the two lenses converge. In fact, the Bukowski and Panavision units make specific provision to converge the lenses, not only horizontally – standard practice for most stereoscopic systems – but also vertically, so that there would be little to no parallax at the subject of chief interest onscreen.

Bukowski’s iteration put two optical wedges at the front of the device to effect horizontal convergence. Two matched prime lenses, designated “receiving lenses”, captured the left and right points of view. Light from these primes was conveyed through matching field lenses inside the box-like body of the device. A pair of first-surface mirrors behind the field lenses shifted the right-eye image about 3.25cm (1.28 in) to the left, to match the centerline of the film in the camera. A similar pair of first-surface mirrors, just underneath the first pair, shifted the left-eye image about 3.25cm (1.28 in) to the right. The image pairs, now “stacked” vertically, were then conveyed through a single relay lens onto the film (Bukowski 1984). The center-to-center distance of the stereopairs in this design was 0.374 in (9.5mm) (Anon., 1985). Because the resultant images reached the film in an unconventional orientation (rotated through 180 degrees), it was necessary to mount the camera upside down in relation to the lens unit. Otherwise, optical printing would be required to make rectified, projectable prints.

The Panavision redesign replaced two of the four first-surface mirrors with a pair of Amici prisms, with the result that the images reached the film in the conventional orientation and the camera could be used right-way up. In this Panavision redesign, the left-eye lens is now higher than the right-eye lens. Instead of wedge prisms, the Panavision redesign relies on precise horizontal offsets of the right prime lens to effect desired changes in stereoscopic convergence.

The periphery of the combined stereo pairs measures approximately 0.870 in x 0.735 in (22.1mm x 18.67mm), with a “septum” separating the left and right images measuring approximately 30 mils (0.762mm) (Fazekas, 1985). The precise center-to-center distance in Optimax III stereopairs after the Panavision redesign is an open question. The Fazekas patent specifies 0.3675 in (9.36mm), while literature from StereoVision International, Inc., who supplied projection optics for all over-and-under 3-D systems, gives a figure of 0.387 in (9.83mm). (Anon 1985). 

Since the Findlay and Bukowski lens units were practically hand-built, and since films made with them were rough-and-tumble exploitation and action flicks, optical components could sometimes jar loose or fall out of alignment. The more than 20 glass surfaces in the Bukowski unit required constant inspection, frequent cleaning and occasional realignment – and this was not always done. Films made using early versions of Optimax III sometimes display chromatic aberration, vignetting, speckling, watermarks, uneven focus, and illumination mismatch between the left and right images, often to a greater degree than may be observed in rival over-and-under format lens systems. The Panavision redesign lessened or eliminated many of these problems.

Bukowski’s upgrade of Findlay’s design offered a 32mm prime lens, with add-on optics to achieve equivalent focal lengths 16, 24 and 64mm. (Bukowski, 2006a; Bukowski, 1984). The Panavision upgrade offered five pairs of prime lenses: 16, 25, 35, 50 and 85mm. (Fazekas, 1985).

The Bukowski and Panavision units offered a small variation in interaxial – that is, the separation between the optical axes of the prime lenses. The minimum separation was 2.031 in (51.59mm), while the maximum was 2.875 in (73.03mm). While stereographers generally advocated a smaller interaxial for closer camera-to-subject distances, in order to achieve more comfortable viewing results, one source somewhat ambiguously states that the interaxial for Optimax III varied according to focal length, suggesting that optical/mechanical limitations of the system dictated the interaxial distance (Williams, 1983). But Bukowski’s patent in particular states that the range of interaxials is meant as a feature to enhance the system’s flexibility in use.

Observed in an upright position, with the soundtrack (guided) edge of the film at left, the left-eye image is on top, the right-eye image below. Raw scans from the film Dynasty display no obvious indexing system (i.e., gate notching, or special inter-perforation markings) to distinguish the left- and right-eye views.