Nikonos RS R-UW AF 13mm f2.8 lens – adaptation to modern digital Nikon cameras in a SEACAM housing and review
(May 2014)
The underwater Nikonos RS R-UW AF 13mm f2.8 Fisheye lens was, in the days of film cameras, known as the concept of optical excellence in underwater wide-angle photography. Today, in the era of modern digital cameras, no one considers to use the out-dated film Nikonos RS any more. Nikon, unfortunately, shows absolutely no interest in producing a digital version of the once famous Nikonos RS, thus making the use of these excellent underwater lenses obsolete. The underwater RS system was astronomically expensive so buyers were scarce and Nikon consequently, already after 4 years, due to economic reasons, withdrew it from its production program in 1996.
The adaptation of the underwater Nikonos RS R-UW 13mm f2.8 lens to function with modern digital Nikon camera, has already been described in the recent past, but the solution employed a special adapter which connects three different units to each other: the lens, the camera and the underwater housing. Here, however, I see the following issues:
- The Nikonos lenses were designed for rugged underwater cameras. By adding an adapter to the system, there is a possibility that a hydrostatic pressure could affect a much more delicate modern digital camera and could thereby hinder its operation or even damage it.
- In the case of minor defects on the O-ring between the adapter and the lens, the adapter would be in a function of a funnel, which means it would be leading water directly to the camera and would consequently destroy it. In the case of minor defects on the O-ring of the dome port, only a little water would be accumulated at the bottom of the underwater housing and would thus keep the camera dry and undamaged.
- Mounting the lens and the adapter is more time consuming and more complicated than the installation of conventional ports.
- In the case of SEACAM underwater housing (for D800), zoom gears, manual focus gears, and lens release levers should be removed in order to insert the adapter. This means that some important housing functions would be lost. Removing the RS lens from the camera would be particularly awkward as the adapter prevents finger access to the lens release button, which would therefore make it necessary to open the housing back-cover and press the button from behind.
For this reason I decided to tackle the project from the beginning and with a completely new approach.
The Project
First of all, I looked for the Nikon original patent (I've found the US version, which is filed under U.S. 5579169) and studied the structure of the lens. It should be highlighted that the patent was filed only on 12/09/1994, entering in force on 26/11/1996, at a time when the lens has already been withdrawn from the production line. Another interesting fact is that the patent describes three different lenses:
| Focal length: | Aperture: | Viewing Angle: |
|---|---|---|
| 12,32 mm | 2,87 | 170,6° |
| 12,49 mm | 2,85 | 180,0° |
| 12,10 mm | 2,85 | 166,8° |
The lens we are talking about (and the only one that came into regular production) is the first one.
The scheme from the Nikon's original patent
It is necessary to clarify Nikon's nomenclature in the expression “focal length” of underwater lenses. The term comes already from the times of the French camera Calypso (invented by Jean de Wouters), for which Nikon purchased a license in 1963 and had begun to produce its Nikonos series. Then, the original 35 mm amphibian lens had approximately the same underwater viewing angle as the 50 mm landlens had on surface. When Nikon produced the first underwater-only lens (28mm f/3.5), it was marked "28mm", although the underwater viewing angle was similar to the 35 mm land lens. To obtain the real focal length, which gives us the right information of the viewing angle, the nominal focal length must be multiplied by a factor of 1.33 (the refractive index of water). Since this nomenclature was implemented by Nikon for the RS series lenses as well, the real focal lengths are as follows: 17 mm (for 13/2.8), 27-47 mm (for 20-35/2.8), 37 mm (for 28/2.8) and 67 mm (for 50/2.8).
I purchased my lens in Italy – today these lenses are quite rare and expensive. As I dismantled it I realised that the entire internal group of lenses, when focusing, is moved back and forth with respect to the camera and the first lens element, which acts like a mini dome port. The latter remains static at all times with respect to the camera, exactly the same as the dome port when the camera is in an underwater housing.
That is when I first got the idea to split this lens into two components! On the basis of the patent data and by measuring the internal dimensions of the lens housing I calculated the distance between the front lens element and the inner group of lenses. I constructed an aluminium ring with a SEACAM thread for the front lens element. I even used the original Nikonos plastic lens hood and a ring around the front lens element with a serial number of the lens. I also constructed a new housing for the internal group of lenses. As a result, I got a new lens and a dome port, which operate properly only when they are placed together.
This was then followed by electronic lens conversion. There are several different options on how to convince a lens and a camera to recognise each other. One possible solution is to replace the electronic circuit with a circuit of a similar but
more modern lens. Of course, this is no easy task, so I was forced to find out certain tricks before everything started to work. The camera eventually recognised the lens and all functions worked perfectly (autofocus, aperture stopping down,
all four exposure modes and all three metering modes), at least they seemed to.
Benefits of my version of adapting the lens are mainly the following:
- There is no need to remove some parts from the underwater housing and thereby functionally limit it.
- Installing and removing the lens is much simpler.
- The hydrostatic pressure can't have impact on the camera in any way and thereby can't affect its performance or even damage it.
- There is practically no depth limit. Since the front lens element is relatively small and very thick (at the thinnest point in the middle is 8 mm thick), it would withstand a depth of several hundred meters.
Since I personally had no commercial interest in this project – though it would also be a pity that other photographers wouldn't benefit from it – I presented my idea to Harald Hordosch, the owner of SEACAM. We have known each other for almost 20 years and have also collaborated several times in the past. He was very enthusiastic about the idea and soon began working on this project. The result is a truly superior product, like all SEACAM products are.
My original prototype is on the left and the final SEACAM product is on the right
SEACAM's additional optimisation
Since May 2013, I have been using the adapted lens with my D800 and I’ve thoroughly tested it under different conditions. I have used it in the Red Sea, the Bahamas, Indonesia and Croatia. In Croatia I carried out a great number of trimix dives and the deepest picture was taken at the depth of exactly 100 meters! On all my dives I have used a regular SEACAM D800 housing with the S10 viewfinder and the SEACAM 150 Offshore TTL flash (tested up to 160 m).
Despite the fact that I pointed out at the beginning that all lens functions work perfectly, it was not quite so. I noticed one interesting phenomenon: if a light meter was obeyed, the resulted image was one to two f-stops overexposed like the aperture would not be stopped down sufficiently. There were also two additional phenomena noticed: images taken even at f/22 were not soft (as they should be due to the diffraction problem) and depth of field was unusually shallow.
I presented the issue to SEACAM's expert for electronics. He discovered the mechanical incompatibility between the original Nikonos RS lens and Nikon land cameras for stopping down the aperture. A Nikonos lens requires a longer shift of aperture lever compared to the Nikon's land lenses in order to stop down the aperture sufficiently. An f/2.8 aperture is in fact an f/2.8 aperture, but when the camera is trying to stop down the lens to f/22, the lens is in reality stopped down only to f/14, which is entirely consistent with my observations. In fact, stopping down the RS lens to f/22 would require us to set f/45 value on the Nikon land camera, which is impossible since the lens provides the camera information that its aperture range is from f/2.8 to f/22. A SEACAM electronics expert was able to brilliantly reprogram the electronic circuit in the RS lens so that the lens "thinks" its aperture range is from f/2.8 to f/45, thus giving the camera this false information. If we want to stop down to the effective f/22, the value on the camera must be set to f/45. The benefit of stopping down the lens to the smallest aperture (f/22) is of particular importance in extreme close-up shots, to achieve the maximum depth of field.
The table of settings on the camera and the effective aperture values:
| Camera setings | Effective aperture |
|---|---|
| 2,8 | 2,8 |
| 3,2 | 3,2 |
| 3,5 – 4 | 3,5 |
| 4,5 | 4,0 |
| 5 | 4,5 |
| 5,6 – 6,3 | 5,0 |
| 7,1 | 5,6 |
| 8 | 6,3 |
| 9 – 10 | 7,1 |
| 11 | 8,0 |
| 13 | 9,0 |
| 14 – 16 | 10,0 |
| 18 | 11,0 |
| 20 | 13,0 |
| 22 – 25 | 14,0 |
| 29 | 16,0 |
| 32 | 18,0 |
| 36 | 20,0 |
| 45 | 22,0 |
The image shows an example of an EXIF file when the lens was set to f/45 aperture (effective f/22).
Optical results (testing in the swimming pool)
The question being addressed the most was, is it all worthwhile? How will a twenty years old optical design, without all the modern optical advantages (like aspherical lenses, ED glass, Nano-coating etc.) behave on a modern high-resolution digital sensor such as in D800?
I first tested the lens with my D800 in the swimming pool by shooting my underwater test card and then in the Red Sea in real life conditions. Unfortunately, this was at the time when I was not yet aware of the before mentioned aperture issue. So I basically tested the lens only to f/14 instead of f/22. For this reason, I will not present the quantitative results of my tests.
In general, I noted with great pleasure that the results of optical quality of Nikonos RS R-UW AF 13mm f2.8 lens are very good. In the swimming pool I also compared it to the land 16mm f2.8 Nikon Fisheye lens in combination with a big SEACAM Superdome (SD), which gives the best optical results, according to my previous tests (better than with Fisheyedome (FD)).
The optical quality of this lens is great for the most image parts and degrades only in the extreme corners. A large area of the image (at least 75% of the area) is sharp even when the lens is wide open. The f/2.8 to f/5.6 apertures can be used without fear when there are no important details in the image corners. The f/8 aperture (effective f/6.3) represents the limit where a lens provides sufficient sharpness throughout the image area. An ideal range of sharpness, where the whole image is razor sharp from corner to corner is f/11 (effective f/8) to f/22 (effective f/14) and most probably also beyond, until the diffraction problems start to emerge.
Section A:
Section B:
Section C:
Section D:
The underwater RS lens is also incredibly well corrected for chromatic aberration (see comparative images below). Pictures were taken in RAW format and developed in Lightroom without chromatic aberration correction. This is in a way logical because the lens was developed in the film days when there was no digital manipulation of the already captured images available…
The lens provides close-up focusing, as shown below, but unfortunately still not as close as the DX 10.5mm fisheye lens (using a mini dome port). It should also be noted that the depth of field is significantly lower due to the FX-format (full frame) sensor. In case of close-up focusing, the image background is quite unsharp, even if taken at f/22 (effective f/14).
Finally, it should be mentioned that this lens does not have a profile in Lightroom. According to the optical characteristics, it is most similar to the land Nikon 16mm f2.8 Fisheye lens, but it is still another lens with different optical characteristics. A 16 mm land lens has a viewing angle of 180° and a 13 mm (effective 17 mm!) underwater lens has 170°. If in Lightroom, a 16 mm profile is used to settle curved lines a.k.a. "defishing", the lines would bend slightly outwards (pincushion distortion), as shown in the image below.
In a few words: According to my experience in the swimming pool, the underwater Nikonos RS R-UW AF 13mm f2.8 lens provides the following advantages:
- Excellent sharpness in full FX format, which is incomparable to any land lens in combination with a big dome port.
- A small front element ("dome port") enables a close-up focusing, but again with better optical quality compared to the one of a land fisheye lens behind big port dome (such as SEACAM Superdome),
- Practical absence of chromatic aberration.
There are also some drawbacks, but in general minor in comparison to the advantages:
- Reduced sharpness in the image corners at wide open apertures (f/2.8 to f/5.6).
- A close focusing characteristic of the lens is good but not as excellent as the land DX Fisheye 10.5mm f/2.8 lens.
- Reduced depth of field due to FX format, which can sometimes be disturbing in extreme close-up photography.
- Overexposed images at automatic exposure mode (without taking into account aperture corrections).
- The lens does not have a profile in Lightroom.
- No possibility of over-under shots (on the surface this underwaterlens focuses very closely only and is sharp only in the image centre).
Testing in the sea (prior to SEACAM aperture correction)
After the "laboratory" testing was carried out in the swimming pool, I started testing the R-UW AF 13mm f2.8 lens with Nikon D800 in SEACAM housing in real life conditions. I was shooting in the Red Sea, the Bahamas, Indonesia, and in the Adriatic Sea. In the Red Sea, where conditions for underwater photography are very good,I endeavoured for maximum quality. In the Adriatic Sea, I tested it mainly in extreme depths.
When I reviewed my Red Sea photos on a large (30") computer monitor, I was initially a little disappointed since the sharpness in the extreme image corners was not as good as in my swimming pool shots. Fortunately, I quickly discovered the reason for this – see explanation below.
Fisheye lenses, like the rest of the lenses, focus on flat surface and not on spherical surface. In the swimming pool, where I photographed more or less flat objects (test card), the lens perfectly accomplished its mission. In real life, however,
when we shoot underwater with fisheye lenses, the objects are all around us and in particular those in extreme image corners are mostly before the focus plane (i.e. too close). Since this lens provides relatively small depth of field, objects
lying outside the focus plane are unsharp. The reason for image corners not being so sharp is not in poor optical design of the lens, but in the non-ideally positioned objects.
The image below graphically explains the problem of blurred corners.
The image below (ISO 400, f/22 (effective f/14), 1/80) shows that fishes (in the background) in the upper image corners are much sharper than the rocks and corals in the lower image corners which are too close. It can also be noticed that the top of the left soft coral is quite blurred (because it is too close) but fishes in the background behind it are sharp.
The picture below was taken in the Adriatic Sea at a depth of 95 m (ISO 800, f/18 (effective f/11, 1/13 s).
The picture below shows how very small the depth of field in close-up shooting is (ISO 200, f/16 (effective f/10), 1/100), where even the entire main object is not completely sharp.
However, it should be remembered that due to its small pixels, the D800 apparently provides more shallow depth of field than cameras with lower resolution. Consequently, a picture taken with the D4, for example, would appear sharper.
Conclusion
Of course we cannot expect the best optical quality from a twenty-year-old lens, however its sharpness on high-resolution cameras such as the D800, is satisfactory enough. I can imagine that in the film days it was a true optical pearl!
Personally, I would say that the proportion between the image sharpness in the middle and in the corners is comparable to a Nikon DX Fisheye 10.5mm f/2.8 lens in combination with a big dome port (such as SEACAM Superdome), which is according
to my tests the best combination of any land wide-angle lens and dome port. There are two major benefits of this 13mm underwater lens over the DX 10.5mm lens behind big dome: firstly, the image is taken on the full 35 mm format with all the
benefits thereof instead of on a small cropped sensor, and secondly, instead of a big and clumsy dome port, a small "dome port" is used. In any case, the R-UW AF 13mm f2.8 lens, in all respects, provides better optical results than
any full frame fisheye lens in combination with any dome port.