III: The Nikkor Lens

HAL’s eye

Finally we get to the glowing glass eye: the key element to HAL’s appearance. Various people over the years have proposed theories as to what kind of lens the computer used to peer out into the world. Some suggested they were custom-made. Oddly, for a while fans thought that a cheap Kenko fisheye lens adapter was the thing. Odd because not only do I not see Kubrick having used a low-end consumer product like this, but the Kenko lenses were far too small!

Evidence now clearly shows that lightly modified Nikkor 8mm f/8 circular fisheye lenses, sold for 35mm still film photography, were used to make the screen-used HAL 9000 faceplates.

NIKKOR or Nikon?

To start, yes, it’s technically anachronistic and incorrect to refer to these things as Nikon lenses. The Japanese company we know today as Nikon was still called Nippon Kōgaku K. K. back when 2001 was filmed – it was their line of 35mm still film cameras that was called “Nikon F”. The firm has branded all their lenses under the “Nikkor” name since 1932.

However, the company itself changed names in 1988 to Nikon Corporation, in order to capitalize on the brand recognition of the camera. So most people generally refer to these lenses as “Nikon” optics, even though they were officially branded “Nikkor”.

The fisheye

Regardless of what they were called, the 8mm f/8 fisheyes were manufactured in Japan between 1962 and 1965, and thus would have been available for the film’s production dates – 2001’s live action sequences were shot in England between 1966 and 1967.

Nikkor 8mm f/8 lenses are 82mm in diameter at the widest point – the knurled grip mid-barrel – and that size corresponds exactly to the 4K photo when dimensions are scaled to the Ordway/Johnson blueprints. Examination of screengrabs indicates that there was no variation in lens size across the seven HAL faceplates seen in the film.

That’s encouraging, but the key evidence comes from both screencaps and still photos that show the side of the lens. No photos exist, that I know of, that are sharp enough to reveal the actual text on the lenses in legibly incontrovertible detail. However it is possible to make out white blobs where the writing is located. By comparing these blobs to the position of text on actual Nikkor 8mm f/8 lenses, it’s possible to show a direct correspondence. Amadeus Prokopiak used this method with a black and white still photo taken in the pod bay set during the “testing AE-35 unit” sequence. He shows that the HAL lens in that particular scene matches a Nikkor 8mm.

I decided to see what we can see in the actual film itself.

This is a 4K screencap that I enlarged and sharpened, from the testing sequence in the pod bay. I also rotated it 180°. Note how the blobs on the left correspond clearly to “1:8 f=8mm”.

Another useful data point is that the Nikkor 8mm f/8 had a very unusual conical metal lenscap. It’s clearly visible in this pod bay photograph by Clive Arrowsmith.

Could other lenses have been used on set? Yes, it’s entirely possible. For example, the Dennis Gilliam replica shown at exhibitions uses a Nikkor 7.5mm f/5.6 fisheye lens, which is exactly the same size, and which has the same lens cap design. It’s chronologically plausible since that lens was introduced in 1966, about the time that filming began in England. Maybe one of those was used. However, I have not seen any photos that show a 7.5mm lens. Whereas we have photographic evidence that the centrifuge and pod bay faceplates both contained Nikon 8mm f/8 lenses.

That would be cool if we could read the actual serial number of the lens used in this scene! The first three white blobs are “Nippon Kogaku Japan”. The last blob is “No.” followed by the serial.

Oh, what are those numbers? Sadly, even extensive digital processing does not reveal the answer. From what we know about Nikkor 8mm f/8 serial numbers that’s either “No.88” or “No.89” at the start there, followed by three other digits. Thus it couldn’t be a Nikkor 7.5mm lens, which had serial numbers starting with 75. Nikon serial numbers are not strictly chronological.

The HAL lens seen on the front of the EVA pod during the murder scene is also a Nikon lens, but unfortunately the angle of the shot means we can only see a few letters. Specifically we can make out “aku Japan No”, which corresponds exactly to the lettering of both Nikkor 8mm f/8 and 7.5mm f/5.6 lenses. If only the lens had been rotated clockwise a few degrees!

Another piece of information supporting the Nikkor 8mm idea is that it’s clear that the lens barrel consisted of two separate thin-walled cylinders, one within the other. This odd design is because the inner cylinder is the section of the barrel that retains the primary lens element, and the gap between the two cylinders is internally threaded on both sides and used for attaching the sturdy metal lens cap. This complex setup is quite unusual.

Finally it’s worth pointing out that the retrofocal Nikkor 8mm f/2.8 was not used in the film, as some people have suggested. First, it looks quite different, and second, it came out too late, in 1970. Kubrick did not, so far as we know, possess a time machine.

HAL’s glowing light

If you’re a photographer using a fisheye lens you won’t be able to put a filter on the end, because of the huge bulbous piece of glass. So Nikon put a rotating filter disc deep into the body of the lens itself. You turn a ring to swap the filters out. Fortuitously one of the filters is a standard red one for high-contrast black and white photography. Which made it easy for the filmmakers to stick a normal white light bulb behind the lens and get HAL’s famous eerie red glow.

Of course, the idea of a glowing lens really makes no sense whatsoever. No light-collecting device like a camera is going to emit light through its objective lens: it wouldn’t be able to see anything! As a photographer Kubrick would have been totally aware of this. But the addition of a simple light source makes HAL look more alive, somehow.

Incidentally, people sometimes put red LEDs into the HAL replicas, which doesn’t give quite the right look. If you check the photo you’ll see that the very centre of the HAL eye is actually yellow, and the glow around it is red. This is because the original props had steadily lit tungsten light bulbs at the back, which produce a warm white light. The light overpowered the thin filter in the middle, resulting in a yellower colour, but further out the red colour was more apparent.

Some Modifications Required

Now one thing I discovered when I made a replica of the HAL 9000 faceplate using an original Nikkor lens is that simply sticking a light source at the back of the lens doesn’t give you the actual look of the lens as it appears in much of the film.

Normally when you shine a light through the back of a fisheye you end up with a really tiny dot projected onto the main lens element (the outer curved glass surface). This dot isn't very easy to see unless you're looking into the lens straight on. The reason is to do with optical design - fisheyes contain a bunch of converging lenses that focus the image down to a small piece of film.

Now, HAL actually does look like this much of the time in the movie – look closely at any long shot of a faceplate in a panel and the red light will usually be quite small. 

A very small red dot!

But we’re used to seeing the red light fill up a big portion of the lens, because in closeup views that’s what we see in the movie.

I did a lot of experimenting with a real Nikkor lens, and it turns out that to get that expanse of glowing light you have to dismantle the lens and take a bunch of stuff out of it.

As a maximum f/8 lens it has a very small aperture even when wide open. I found that the filter disc, the aperture diaphragm assembly, and all the glass lens elements except for the primary and secondary all had to go. They're all either threaded or screwed in place, so it's possible to remove the components and store them carefully for future reassembly. 

The result? HAL's baleful crimson eye! The photo below is of my replica HAL, as part of the Kubrick Exhibition in Madrid.

What is a fisheye lens anyway?

Camera lenses work by projecting circles of an image onto the surface of the film or the image sensor. Most lenses have image circles that are bigger than the area that records the image. That way the scene fills the whole finished photo. You never see the edges of the projected circle, as it’s cut off.

Nearly all camera lenses are rectilinear, which means that they record a scene so that, for example, square objects look square in the finished photo. In other words, lines that don’t pass through the very centre of the image are rendered as straight.

But fisheye lenses are different. Only lines that pass through the centre of the image are straight. All others are recorded as curves, and the further a line is away from the centre, the more it’s curved. This is “barrel distortion”, which means that squares are rendered as bulging, convex, objects. Fisheyes are extremely wide angle lenses capable of taking in huge swathes of a scene, but at the cost of looking bulgy and trippy. Think of the peephole viewers built into front doors.

This is a photo I took of London’s Natural History Museum using a 16mm full-frame fisheye lens. Note how lines that should be straight – the handrails to the left and the stone wall to the right – appear to be bulging outwards.

Fisheyes are of two basic types. Full-frame fisheyes have image circles that are larger than the surface of the film or sensor. This means they cover the entire picture. They also take in 180° of a scene if you measure across the diagonal of the photograph.

By contrast, circular fisheyes have image circles that are almost exactly the size of the film or sensor. They thus show round images with black corners. They take in 180° of a scene measured across the short edge of the photograph.

This is a photo of the Eiffel Tower that I took using an 8mm circular fisheye lens – though a Zenitar lens, not a Nikkor. Note how the circle almost fills the whole frame, and how the lens gives a full 180° view of the scene. Such lenses are difficult to use as it’s so easy to get your feet in the shot!

A circular fisheye is therefore the perfect lens for an all-seeing AI like HAL. One lens on a wall would give full coverage of the room (except for stuff obscured by furniture, etc) since you get 180° of coverage. Of course, such a visual system would require a lot of processing power to interpret shapes moving across such an extreme curved area, especially towards the edges of the frame. Though that is making a load of assumptions. Maybe in the 2001 future they have sphere-shaped image sensors that record uncorrected data from a full fisheye with ease!

Fisheyes aren’t named after the bulging eyes of fish, surprisingly enough. The name was coined by American physicist R.W. Wood. It refers to the view that a fish would see looking up at the sky from water.

For more information on camera lenses, check out my book on the subject – cunningly named “The Lens”.

The Nikkor’s optics

The Nikkor 8mm f/8 is an interesting lens. It’s a circular fisheye, and so projects round images with black areas in the corners. It’s sometimes described as the first production fisheye made for 35mm cameras. Before its release only specialized lenses were available to achieve its look. Stanley Kubrick, who started out in still photography and maintained a keen interest in the subject throughout his life and loved Nikon products, would have been well aware of the lens.

A diversion into optics – or, what’s up with the lens’ weird-ass back tube?

The lens is an early fisheye design for Nikon, and is not retrofocal. What does that mean? Well, the focal length of a lens is a key factor in determining how much of a scene the lens can take in. Wide angle lenses have short focal lengths, but that means that the back lens element has to be located physically close to the film or sensor surface. The problem is that SLR cameras contain shutter and mirror mechanisms that sit really close to the image plane. Collision time!

Modern wide angle lenses solve this by being optically complex “reverse telephoto” or “retrofocal” designs which place the “rear nodal point” up inside the lens assembly itself. There is thus no correspondence between the rear nodal point’s distance from the film/sensor surface, and the lens focal length. This allows the lens to have adequate internal clearance for the mirror.

Unfortunately the non-retrofocal Nikon 8mm f/8 has to have this long protruding tube that sticks out past the lens mount, and goes deep into the camera body. In order to use it on a mirror-equipped SLR you have to flip and lock the mirror into the up position – assuming it can, and not all SLRs can. If you don’t you’ll smash the mirror when you try to use it. In addition to limiting compatibility it also means you can’t look through the camera’s viewfinder to focus and compose – an external finder is required.

This highly impractical design was later superseded by retrofocus fisheye lenses, such as the popular Nikon 8mm f/2.8 fisheye, which could be used more conveniently on SLRs. Such modern lenses don’t have this tube sticking out past the mount.

This is a cutaway drawing of the Nikkor 7.5mm fisheye lens. I’ve tinted the sections of the diagram that indicate glass – this lens was made of 9 separate pieces of glass (lens elements) that were arranged in 6 groups. The 8mm lens was similar, but its optical formula differed.

I believe this is the optical diagram for the Nikon 8mm f/8 lens, which had 9 elements in 5 groups.

The hexagonal aperture diaphragm can be seen briefly in the movie (the Poole murder closeup), so this component was left in place, wide open (at f/8 of course) for the scenes when the lens wasn't modified to produce a larger area of red light. Unlike most fisheye lenses sold today with SLRs, the Nikkor 8mm f/8 is fixed focus, and has no focus adjustment ring.

The lens Coatings

Being a lens from the early-mid 1960s, the Nikkor f/8’s glass was single-coated, and so it didn’t do a great job at reducing reflections. Modern camera lenses, by contrast, usually have complex transparent multicoatings which yield those characteristic green or red reflections that you see. The Nikkor 8mm f/8 does not have such coloured reflections, and in fact its lack of multicoating is one reason why it’s such a slow (f/8) lens.


II:   Ingredients of the Faceplate

IV:  The Different HALs


Contact

© 2021-23 NK Guy, contact@theageofplastic.info