Best wide angle lens for Nikon D850 in Nauticam housing

[SMY 15]

I love my sigma fisheye but Im looking now for a wide angle lens for my Nikon D850 in Nauticam housing. Which one do you prefer or do you recommend...and with which port?

 

Regards, Sascha

[Kraken de Mabini 202]

With my Nikon D800e in a Subal housing I use two wide angle lenses: a Tokina 10-17 fisheye and a Nikkor 18-35 mm.

For ports I use a small dome, 100 mm dome, and an 18 mm extension. A larger dome might be a good idea, to get a smooth image with no corner cropping.

[Undertow 31]

16-35mm for a number of reasons. Need a big dome 8-9"+ for rectilinear superwides.

[SMY 15]

16-35mm for a number of reasons. Need a big dome 8-9"+ for rectilinear superwides.

 

....I hear very often, that the 16-35 is in favor, why? How about the 12-24 or the 17-35?

[TimG 62]

I'd have though either the Sigma 15mm or the Nikkor 16-35mm. Isn't the 12-24 a DX lens?

[SMY 15]

I'd have though either the Sigma 15mm or the Nikkor 16-35mm. Isn't the 12-24 a DX lens?

 

Sorry, you‘re absolutely right! I meant the 14-24 FX lens......

[TimG 62]

Sorry, you‘re absolutely right! I meant the 14-24 FX lens......

 

No worries. I read on WP some time ago that the 14-24 was very hard to house successfully and the 16-35 worked better. Worth a search for it.

[trimix125 73]

Hi,
the Sigma 12 - 24mm is a FX lens since the first generation, as far as i know......
A glas monster, but was the first ultra wide (weitwinkel) angle for film / FX.
And is f4 now in the Art series.
Regards,
Wolfgang

[TimG 62]

Sacha

 

I found this on http://www.borutfurlan.com/articles/dxfx-en.htmlabout the use of the 16-35 and 14-24

 

 

Dome port theory

Dome ports are the most common correction system used in underwater wide-angle photography. The dome port is defined by two spherical surfaces (outer and inner) having an identical centre of curvature (centre of sphere) and therefore forming a concentric dome. The difference of curvature radiuses of both surfaces equals the dome port thickness, which must be completely uniform.

The basic operating principle of a dome port is bringing the light rays into the lens in such a way that there is no refraction on the water-glass-air boundaries. This removes all CA, distortions and magnifications as the main optic problems occurring when the photos are taken from behind a flat port. This is only possible if the rays are entering the glass (almost) perpendicularly. Since all lines going through the centre of a sphere intersect the spherical surface at the right angle, the basic conditions for the successful correction with a dome port are placing the lens exactly in the centre of the sphere and perfectly aligning the lens' optical axis with the dome's optical axis. The light rays enter the glass approximately perpendicularly and therefore do not refract, retaining the lens angle of view.

Consequentially, only zoom lenses that don't change its physical size significantly while zooming are suitable for photography with a dome port. Among the Nikon FX lenses, the 16-35mm f/4 is much more suitable than the "legendary" 14-24mm f/2.8, the latter changing the position of its front optical element strongly while zooming.

[SMY 15]

Brilliant! Thanks a lot Tim!

 

So the race is now only between 16-35 and 17-35....

 

Best regards, Sascha

[Undertow 31]

....I hear very often, that the 16-35 is in favor, why? How about the 12-24 or the 17-35?

 

Sorry, you‘re absolutely right! I meant the 14-24 FX lens......

 

I read a good bit when deciding 16-35 vs 14-24. 17-35 wasn't a consideration (price, less wide).

 

Many people struggled to get acceptable results in corners (challenge of rectilinear superwides) with the 14-24mm. Its much bigger, heavier, can't take filters, no VR and costs nearly twice the price. I believe some housing brands can't fit it too. It is a gorgeous lens though.

 

There are times I wish I had 14mm (or 2.8) but there are MORE times I'm glad to have 35mm underwater. Sharks, dolphins, turtles, squid etc.

 

Topside I've shot some lovely landscapes with a polarizer on the 16-35 (internet will say never use a pol on superwide, but works great in certain situations).

 

Cheers,

 

 

Chris

[adamhanlon 0]

There is quite a lot of information on this forum about wide angle choices.

 

A lens' performance on land is not a reliable indicator of how it will perform behind a dome port. The 14-24mm vs 16-35mm debate is a case in point. The 14-24mm is a superior lens, but does not works well behind a dome. Whether a lens is f/2.8 or f/4 (or f/5.6) has no relevance on FX as the shallow depth of field at these apertures means that sharp corners are impossible to achieve. I guess if all you are doing is shooting big animals in the blue, this is not a major consideration, but for everything else it does! Yes, you could crop the softness out, but this would also negate the advantages of a wide FOV!

 

BTW the 17-35mm's corners are not great on land....

 

Rectilinear wide angle with the D850 needs more examination though. For me, images produced by the 16-35mm with a 9" dome on the D810 was acceptable at apertures on around f/12 or so. I would expect the additional resolution to negatively impact on this, to the extent that the 16-35mm may no longer be a viable option.

 

My feeling is that water contact lenses like the Nauticam WACP will probably be the best options for rectilinear wide angle with it.

 

Adam

[ianmarsh 38]

One other thought...

 

I have some experience with the Nikonos RS 20-35mm Seacam conversion. Initially, it had it's detractors who didn't feel it was that useful. 20mm was just not wide enough. (Stephen Frink seemed to do OK with it in the '90's:))

 

There is some truth in that, but for the right subjects, I was very happy with the results. The small(ish) form factor and razor sharp corners were excellent. Also, unlike the 13mm RS conversion, the original lens is NOT selling for $3500+ USD. It is much more "affordable" than it's upscale cousin.

 

Of course, this also assumes a Seacam housing. The zoom gear is integral with the lens, and despite numerous attempts, I was not able to make it work in a Nauticam housing, which the 13mm will.

 

So, now I have a Seacam housing...

 

I have now pared my UW lens collection down to the Nikonos RS 13mm, RS 20-35mm, Nikonos 15mm on Sony, and Nikon 105G for macro. The only thing I can't do is under/overs.

 

I am completely satisfied with my setup. Now, if I were just a better photographer...

 

ian

Edited January 9, 2018 by ianmarsh

[ChrisRoss 150]

 

Consequentially, only zoom lenses that don't change its physical size significantly while zooming are suitable for photography with a dome port. Among the Nikon FX lenses, the 16-35mm f/4 is much more suitable than the "legendary" 14-24mm f/2.8, the latter changing the position of its front optical element strongly while zooming.

 

 

I think this needs a rider the issue is movement of the entrance pupil while zooming whether or not it's a problem is going to depend to some extent what focal length range you are considering. For example the Olympus 12-40 does quite well at all focal lengths but it's only wide in the very bottom end of the zoom range (2x crop) and it gets shorter marginally from 12- 14 before getting longer and its entrance pupil doesn't seem to move a lot when zoomed. The 14-24 is very wide and even at the long end is still a bit wide.

 

Regarding the need for better lenses for the D850, if you believe the premise behind the WACP wet lens, that the limiting factor is the dome port and a 28-70 kit zoom can do better than a superzoom behind a dome the natural conclusion is that the extra resolution is empty pixels and underwater a better lens can't deliver because of the dome interface. This may depend on how well optimised the dome is for the lens in question. which could be up to 10mm out from optimum placement if your chosen housing uses the principle of taking a standard dome and adding extension tubes with fixed sizes like 10, 20, 40 etc. so you are limited by what size extension tubes are available,

[Geo Cloete 8]

I must say I am still a bit baffled as to why the high megapixels of the D850 is considered to be part of the problem when discussing soft corners.
Does the various dome port theories not deal with the curvature of the dome, the way the light rays bends through the dome glass and how wide the lens being used behind the dome are?

To the best of my knowledge, none of the theories state that the light will bend differently depending on the megapixels of the camera.

Alternatively, if considering that the 14-24mm paired with the D850 delivers beautiful results topside, but not underwater surely we can agree that what has changed in the equation is not the megapixel count going from topside to underwater, but the fact that the 14-24mm is now behind a dome port.

[adamhanlon 0]

I disagree with the assertion (and the article excerpt above) that a moving entry pupil is the issue with wide-angle lenses and domes on FX. It may be the problem with some lenses, but others with a moving front optical elements work fine. It is down to how well an individual model of lens copes with the curved virtual image and the depth of field issues that this creates. The only way to know how a specific lens performs is to put it behind a dome port and test it underwater! Land tests etc. have no relevance to this.

 

Ironically, choosing a fast (f/2.8 etc.) lens is also a complete red herring! Typically, large apertures give very small depths of field, which means that corners will be soft.

 

Lens to port distance is crucial too and it will pay to spend time and gather information about the correct extension length. This is not as simple as it sounds either! Not specifically relevant to this discussion, but the optimal length of extension for the venerable Tokina 10-17mm with a few housings is actually 17mm. To my knowledge, no manufacturer supplies a 17mm extension.....So these are inevitably a compromise. In terms of the 16-35mm with Nauticam and 230mm dome, the correct extension seems to be between 60mm and 70mm, with most people opting for the latter. This (and all the other optical) compromise becomes more problematic as we increase the camera resolution....

 

Geo is entirely correct in asserting that the physics of the light passing through the dome, entering the lens and being recorded on the sensor is not affected by pixel count. However, the extra resolution of the D800/810 (and now D850) dramatically show the optical flaws in the lens/dome interface. When we were using 20MP cameras, the level of detail that they recorded meant that, as long as the photographer was careful, the optical flaws in the system we used were not really noticeable. The increase in resolution means that the sensor records more detail and the D800/810 (along with the Canon 5DS/5DSr), really showed that the 16-35m/230 dome combination was not as good optically as we all thought. The D850 increases the level of detail in the images it produces by another 25% (roughly) so any flaws will also be 25% more visible. I have not shot the D850/16-35mm combination, so cannot speak from experience, but I have shot it a a lot with the D810 and am aware of its shortcomings at 36MP.

 

What saves the day, as far as i am concerned are optical correctors that are designed specifically to deal with the water/air/lens interface. These optically correct the curvature of the virtual image, presenting an image to the sensor that is on a parallel plane. Nikonos lenses were examples of this and the 13mm is a great lens, although it is a fisheye so not really germane to this discussion. Ian has pointed out the advantages of the 20-35mm RS lens. For me though, they are not really wide enough.

 

So we are left with the Nauticam WACP. The advantage of this lenses is that they correct the image so that the lens is able to produce an image that is optically the same as one produced in air. None of the standard lenses or cameras that we use are designed to deal with the issues of a curved virtual image. Land photographers (which is where most cameras get used) will never bump into this issue! While we were using 20MP cameras (or DX), we weren't able to see the flaws. The increased details available to us with high resolution FX cameras shows them up.

 

Why is all this important? To be honest, if you are planning to share low res images online, it probably isn't. When we are underwater and shooting, there are so many things that we cannot control. Light varies, subjects are skittish (or absent), waves and surge make maintaining a stable camera platform a challenge, air supplies are finite and so on. One thing that we can control is our equipment choices. In order to give ourselves step best chance of obtaining the best image possible, we should not compromise. Choosing a camera like a D850 means that you are serious about image quality. To compromise this by using optics that are sub optimal seems (to me) to be a bad idea!

 

Adam

[Undertow 31]

Personally I don't see what people say about higher resolution cameras showing 'flaws' to some point of damaging the image, though I keep reading it.

 

I shot with 10-12mp cameras for many years, then jumped to a D800 (topside) and only saw vastly improved technical image quality, namely better resolution and dynamic range.

 

I upgraded my UW housed D700 for a D810 a year and a half ago and again, only seen improvements. The resolution jump to the D850 is less significant and I personally doubt I'd see any worse performance from my 16-35 + 9" dome.

 

Ironically, choosing a fast (f/2.8 etc.) lens is also a complete red herring! Typically, large apertures give very small depths of field, which means that corners will be soft.

 

2.8 lenses do have noticably superior autofocus as the AF sensor has a brighter image with which to work.

 

For me, images produced by the 16-35mm with a 9" dome on the D810 was acceptable at apertures on around f/12 or so. I would expect the additional resolution to negatively impact on this, to the extent that the 16-35mm may no longer be a viable option.

 

Here's a shot D810, 9" dome, 16-35 @16mm, 1/200 f/8 ISO800. Overall image quality isn't perfect due to low light (30mins before sunset), but to me the corners are totally acceptable (corner crop brightened +1.5 stops for clarity here).

 

Had another image from this series blown up about 2m across and looks fantastic. I don't think the D810's resolution makes the imperfect IQ worse at all.

 

 

Cheers,

Chris

Edited January 10, 2018 by Undertow

[adamhanlon 0]

Lovely image Chris,

 

As always, is is hard to make any significant observations from images at screen res, and what is acceptable is very much a personal decision. Based on the crop of the right hand corner above, you can see how it is much softer than the center of the frame. So this is not an IQ issue, but a corner sharpness one!

 

In my opinion, the beautiful lighting of the image combined with the fact that the corners are in shadow means that it is not too noticeable. The distortion is definitely there though. In a similar scene with different lighting it may well be more noticeable and hence detract from the image.

 

For what it is worth, with the D810 and this scene, my inclination would have been to push the ISO up to 1600-2000 and close the aperture down to f/12 or so. The D810 deals pretty well with noise, and LR is very good at getting rid of it, yet corner sharpness is not something you can do anything about except crop it out.

 

AF systems on newer cameras are very capable with slower lenses. Nikon used to use f/5.6 as the cut off for AF. At (lens) apertures larger than that, the phase detect sensors do not have sufficient angle of view so the AF sensor cannot "see" out to the edges. With the D810, Nikon modified the CAM3500TX AF sensor so that some focus positions can now be used with lens apertures between f/5.6 and f/8 and some at f/8.

 

The belief that AF with faster lenses is better as there is more light falling on the AF sensor is (mostly) incorrect. It is to do with the angle that the light falls on the AF sensors. Hence, with the D810 and other newer cameras, there is no issue using slower lenses, beyond the ability of being able to open up the aperture to get sufficient light to expose the image correctly. Given that with rectilinear lenses on FX using a lens at larger apertures does not produce sufficient DOF, it is a non issue for rectilinear wide angle lenses that we use underwater.

 

Adam

[Undertow 31]

Thanks for your thoughts Adam.

 

I was actually referring to both the dome corner sharpness issue and overall imperfect IQ there. To me the dome related corner blur in that image isn't bad at all and wouldn't be distracting even if not in the shadows. If not obvious at screen res, the blur is really only in the lower right quarter of the crop. The left side is at normal image sharpness (which isn't perfect in that image!).

 

Out of curiosity I found another image last night, again low light pre-sunset, where I accidentally shot some images at f/6.3 of turtles over sand/seagrass. The bottom corners were closer to the camera and they're total freaking mush. So it depends.

 

For what it is worth, with the D810 and this scene, my inclination would have been to push the ISO up to 1600-2000 and close the aperture down to f/12 or so. The D810 deals pretty well with noise, and LR is very good at getting rid of it, yet corner sharpness is not something you can do anything about except crop it out.

 

 

It is all about personal preference. I find the corners in that image very acceptable so in retrospect I would not have pushed the ISO as it would sacrifice dynamic range, upon which that image is very reliant. I absolutely agree about noise though.

 

You're right that AF variations for wide lenses underwater are generally insignificant. I have seen a difference but am also totally happy with my f/4 lens there. It is the only lens I own slower than 2.8 and I occasionally bemoan that, but usually just topside.

 

Chris

Edited January 10, 2018 by Undertow

[vbpress 4]

 

BTW the 17-35mm's corners are not great on lamd

 

Adam

Are you sure? Not in my experience. I use 17-35 massively in land, which is a lens that need to be closed by f/8 f/11, anyway I've made a lot of wide aperure shots and the pictures are good, very good. I tried the 16-35 without recognise particular sharpness improvement, inspite of that the AF performane is better and the VR function can help in partiular occasions, but is a less fast lens (f/4) and a little bit bigger. Of course, Nikon users like me are waiting for a modern version of the old 17-35/2.8, but Nikon at that time appears interested to others strange things (180-400/4 Tc a Rolls Roice tele zoom, good luck Nikon)

 

In water using D800/D800E behind Sea&Sea 240 achrylic dome port, I've reached good very good sharpness from corner to center with the cheaper 18-35 Afs closed to f/11 - f/13 adding the correction lens sea&sea CL77 and a couple of extension ring 20+40. But still there is a great lack in the minimun focus distance: 27cm are still too much.

I plan to purchase a nikon 20/1.8 afs and, probably, the newest Sea&Sea 230 glass dome; finally, I hope, I'll have the anphibic rectilinear wide angle camera that I'm looking for.

Edited January 11, 2018 by vbpress

[adamhanlon 0]

Hi Valerio,

 

Sorry, I'm going:

 

I think you have hit the nail on the head with the corner sharpness of the 17-35mm. When discussing corner sharpness on land, the 14-24mm gives pretty sharp corners at f/2.8. Even the inexpensive 18-35 f/3.5-4.5 does pretty sharp corners at f/4.5! As do many other Nikon and Sigma primes. As you say, the 17-35mm does not, unless you close down the aperture to f/8-f/11.

 

The Sea & Sea correction lenses are a very interesting idea and something we haven't discussed here. I have found that they add around a stop to the 16-35mm, when used with a big dome. It will be interesting to find out whether the D850 will pick up the (deliberate) optical distortion caused by them. This definitely needs testing!

 

I tested the Nikon 20mm f/1.8 (http://wetpixel.com/articles/review-nikon-fx-wide-angle-lenses/P3) with the D800 and found it OK at f/11 or so. It was good for big animals in the blue, but I found the 16-35mm f/4 better. This is particularly true when the 16-35mm was used with the Sea &Sea correction lens.

 

The best performance that I have found is actually a legacy lens. The Tokina 17mm f/3.5 is, for some obtuse optical reason, a great performer. Here is an image taken with a 170mm dome at f/6.3:

 

(Nikon D800, Tokina 17mm f/3.5, Zen 170mm dome, two Seacam Seaflash 150 strobes. 1/200@f/6.3, ISO800)

 

Again, I have no idea how it will perform at 46MP. My gut instinct is that it will not be good as it is an old lens that simply was not designed to capture this level of detail.

 

Minimum focus distance is an interesting statistic. Remember that the lens needs to be able to focus on the virtual image created in front of the dome. The virtual image extends from just outside the dome itself (for very close objects), to 3.03 times the radius of the dome (for objects at infinity). So for a wide angle scene with the subject at infinity and a 230mm dome, the lens need to be able to focus 348.45mm (1) in front of the lens. Obviously, if an object is closer than this distance, the lens will need to be able to focus on it.

 

Apologies for going "off the reservation!"

 

Adam

 

 

[Rui_Guerra 6]

 

Consequentially, only zoom lenses that don't change its physical size significantly while zooming are suitable for photography with a dome port. Among the Nikon FX lenses, the 16-35mm f/4 is much more suitable than the "legendary" 14-24mm f/2.8, the latter changing the position of its front optical element strongly while zooming.

 

 

 

Not entirely correct. I have here the Nikon 16-35 f4 and it changes the position of the from element, getting 9mm shorter when going from 16mm to 24mm. From 24 mm to 35 mm it moves out the front element for about 6mm. So there should be other things to consider and not just if it changes or not the front position element while zooming. It also moves a lot (around 30 to 40 mm) the position of the rear element while zooming in from 16 to 35 mm (maybe this compensates?).

 

 

In terms of the 16-35mm with Nauticam and 230mm dome, the correct extension seems to be between 60mm and 70mm, with most people opting for the latter.

 

 

Nauticam recomends a 90 mm extension. Do you have any comparation test shots where you've based your decision about the extension length?

I use an 80mm extension, just because that's what I have here (although image quality is very good, with the normal limitations of FOF and dome ports - using a big 230cm).

[adamhanlon 0]

Hey Rui

 

Nauticam actually recommends either a 70,80 or 90mm extension, depending on which (Nauticam) dome port it is used with...!

 

Seacam recommends a 70mm with their Superdome and a 65mm extension with their WP Wide Port.

 

The construction, shape and fitting of the port itself impact on the correct extension.

 

I have not done any testing with differing extensions with this lens, but have used the recommended ones. It would be interesting to experiment...

 

Adam

[ChrisRoss 150]

I must say I am still a bit baffled as to why the high megapixels of the D850 is considered to be part of the problem when discussing soft corners.

Does the various dome port theories not deal with the curvature of the dome, the way the light rays bends through the dome glass and how wide the lens being used behind the dome are?

 

To the best of my knowledge, none of the theories state that the light will bend differently depending on the megapixels of the camera.

 

Alternatively, if considering that the 14-24mm paired with the D850 delivers beautiful results topside, but not underwater surely we can agree that what has changed in the equation is not the megapixel count going from topside to underwater, but the fact that the 14-24mm is now behind a dome port.

This is ongoing issue in all photo forums above or below water, suddenly you have extra resolution and when you look at the corners at 100% in Photoshop you discover that the image does not look as sharp as the lower resolution older model, therefore the new camera causes poor corners. In fact the corners are the same as they always were and the image will look the same if you downsample to 36 MP or whatever the resolution of the previous model was.

 

The actual situation is you can't take advantage of that extra resolution in the corners because the lens does not resolve that well in the corners.

[vbpress 4]

Hi Adam

Thank you for your precious off-topic, but I still puntualize that the wide angle lenses have a "natural" applications in all focus views, in other words closing the aperture. This is more true behind a dome in uw photography where we have the problem of the corner blur due to the dome out of focus aberrations; we have only two way to reduce the problem 1) increase the dome radius 2) closing the aperture to maximise the deep of field. Well, in my experience f/11 f/16 are abitual aperture in uw wide angle photo shooting.

I remember your experience with Tokina 17/3.5 (surpising experience!) also the Nikon 20/1.8 test. Last year I had problems of out of focus shootig toads just behind the 240 dome using Nikon AF-s 18-35/3.5-4.5 because the amphibians were too much close to the dome for mfd of the lens. This is the reason of my interest for Nikon AF-s 20/1.8 but you have reduced all my enthusiasm

In any case, staying on topic, I believe that the current best wide angle/dome combo for 36mp camera is also the current best combo for 45mp camera.