There are several myths surrounding fisheye lenses when it comes to underwater use.

A quick tour to underwater photography guide usually a good source of information.

Myth Busting

Properties of fisheye lens as per UWPG:

• They usually focus very close
• They are small and light
• When shooting ultra-wide angle, you benefit from a large depth of field
• You can get very close to large subjects, maximizing color and sharpness
• They perform well behind dome ports with good corner sharpness, and they don’t need a diopter
• You usually need at least 2 strobes with good angle of coverage to properly light the entire area.

Some of the above statements are correct in absolute, some are correct but not specific to fisheye lenses and some are just incorrect.

Fisheye lenses usually focus very close -> true for the most recent fisheye lenses, not true for some older models

They are small and light -> Not true. Canon 8-15mm and Nikon 8/15mm are fairly chunky lenses with lots of glass

When shooting ultra-wide angle, you benefit from a large depth of field -> not a property of the fisheye lens but of the focal lens. In fact due to the extreme field of view Fisheye lenses have issues of depth of field.

This is a tea towel shot with a rectilinear lens. Note how sharp the target is at f/5.6

This is the same target at the same distance with the Canon 8-15mm at f/5.6 note how the edges are blurry and the blur starts very near centre.

You need to stop down the lens to f/16 to start getting coverage for the edges.

You can get very close to large subjects, maximizing color and sharpness -> This is a consequence of close working distance and wide field of view however sharpness is another story

As we have seen before fisheye shots at close distance are generally not that sharp especially at the edges.

They perform well behind dome ports with good corner sharpness, and they don’t need a diopter This happens to be true in practice and it is a major benefit for the underwater shooter

We will dive in detail in this topic.

You usually need at least 2 strobes with good angle of coverage to properly light the entire area. Fisheye lenses cover an aspect ratio wider than the format aspect ratio and result in limited vertical angle of coverage. Fisheye lenses are ideal for two strobes except the very far edges.

Canon 8-15mm specifications.

Horizontal field of view: 142 degrees

Vertical field of view: 91 degrees

Let’s ignore the edges and assume we are a one meter.

Horizontal field of view 2*tan(71)=5.8 meters

Vertical field of view 2*tan(45.5)=2.03

Aspect Ratio = 2.85:1

The issue with fisheye lenses is that the frame is really very wide much wider than it is tall. This means some of the edges on the horizontal axis will be normally dark unless you are very very close.

Fisheye lenses and Dome Ports

A dome is simply a lens with a single element that has the property to retain the air field of view of a lens.

A dome is a lens with a lot of field of curvature simply because it is bent.

Using the dome port visualiser we can see that the effect of a dome is to bring the image closer to where it really is.

The net effect of a dome port is to increase the depth of field as infinity focus is reached much sooner.

A dome port has several side effects the main ones are:

• Spherical aberration
• Field of curvature

A fisheye lens works opposite to a dome. The centre of the frame is closer to the lens the edges are further away.

Domes, field of curvature and Fisheye lenses

In order to understant how the barrel distortion works in combination with a dome port and a fisheye lens we can build a small simulation in a light box where the edges of the frame are closer than a flat target.

We can see that despite the edges are quite blurry this image is actually better than our flat target.

At f/11 the image is not perfect but we can see that most details off centre are not looking bad at all.

it is definitely blurry but not as bad as the tea towel as if the way the element are laid out improves the image in the corners.

And this is exactly the point: the items as laid out emulating the curvature of a dome improve the fisheye lens performance.

By f/16 the image is almost all sharp.

One trick is not to focus in the back of the frame but find a middle point this means we can find additional depth of field in front of the target.

Let’s see how this goes. at f/11 we already get some better results.

At f/16 we get some additional improvement but is not as major as the original f/16

Looking at the other areas there are some minor improvements but generally less as we close down the aperture.

In conclusion the layout of the image elements helps the fisheye lens to achieve better image quality this can be futher improve focussing off centre however closing down the aperture results in the best results regardless.

In short we can improve an image at f/11 by shooting off centre in a strategic point to improve depth of field but ultimately aperture plays a bigger role in improving performance of the fisheye lens.

A similar reasoning can be applied to dome size vs closing down the aperture.

We can plot a scenario in the dome simulator tool.

In the starting example our aperture is 4cm to similate our 15mm lens at f/4.

We now reduce the aperture to 2cm which is more or less f/8

And finally to 1cm which is more of less f/14. In reality this is mm not cm but should make you understand that aperture matters more than anything else.

What we can see is that by reducing the aperture the light rays passing through the dome converge and this means stray light is reduced and as consequence spherical aberrations are decreased.

Let’s now introduce dome size which is the equivalent of depth of field in the mix in our light box shooting off centre.

We can see that with a double size dome the converging effect on the light rays is not as significant as the aperture is already small, but nonetheless is present. This is consistent with our f/11 off centre use case.

Finally at aperture completely closed.

Although virtual distance has increased significantly the effect of the large dome on the stray rays is not significant here aperture rules.

What does all of the above mean?

I realise this was a bit geeky.

To summarise a dome has two issues one is spherical aberration for the very shape of the dome. This is mostly cured by closing down the aperture. Dome size has limited effect here unless you shoot wide open and with apertures from f/14 we can see that large dome vs small dome does not really matter.

However when it comes to field of curvature large dome helps the situation but because fisheye lens have barrel distortion and this has a counter effect to dome shape curvature therefore dome size matters much less to a fisheye lens than it would to a rectilinear lens.

Some additional insight in this post. And the summary finding here.

The takeaway message is this: stopping down the aperture improves field curvature and astigmatism somewhat, improves coma a lot, and improves spherical aberration most of all. The sum total of these effects changes our ‘area of best focus’, which is what we photographers really mean when we say ‘field curvature’. 

We could paraphrase this by saying:

A dome port increases depth of field and a fisheye lens, due to barrel distortion, benefits from a dome port. Optical aberrations introduced by the dome are mostly addressed by stopping down the aperture. The size of the dome port does not matter too much when using a fisheye lens and the benefit on aberrations of a much larger size dome is likely to be minimal when we look at that simulator. Focussing appropriately mitigates residual issues of field of curvature of the dome for the fisheye lens.

Underwater proof of concept

I took my Sony A1 with a Canon 8-15mm first and then with a WWL-1 that behaves very much like a fisheye lens.

Let’s have a look at some images shot with Nauticam 140mm dome.

The image above sees two buddies in the frame almost flat with their fins going back in the frame however the result is much better than the lightbox example as result of distance and dome port increasing field of view and adding curvature to bring the fins in.

This however does not resolve all issues if you focus near like in this example focussed on the eye of the croc

Here the eye is close resulting in the tail being blurred this is an effect of close distance and lack of depth of field despite the dome.

More interesting the nose is even more blurred as the dome brings that even close and blurs away due to field of curvature as the focus point is behind.

In this other example instead of focussing on the eye the focus goes mid frame so the fins are still in decent shape even if deep in the frame at f/8.

In order to prove the concept even more I took some props underwater.

First let’s have a look a shot at f/8 with the WWL-1.

As we can see the image is not too bad even in the close area but it is definitely better at f/11

What happens if we position the target off centre?

Contrary to our topside example the situation does not improve by focussing on the edge to further prove the issue here is NOT depth of field.

Here a detail crop the image is still fuzzy despite then focus is right on the spot. Depth of field is not the issue.

And finally we close down the aperture to f/11.

Crop at 100%

So here we can see that the underwater interface provides already for the depth of field but moving the focus at the edges does not have such a good effect.

Why? Because this is likely to do with aberrations of the lens itself as shown in my previous post on the Sony 28-60mm.

The combined 28mm with WWL-1 at f/8 means 20/8-2.5 mm aperture when stopped down to f/11 this becomes small enough to cure aberrations (less than 1cm with reduced field of view is sufficient).

For the same reason ASPC and MFT will be able to shoot at wider aperture not because of depth of field but due to smaller lens aperture.

15mm fisheye at f/14 –> 1.07mm physical aperture

8mm MFT fisheye at f/8 –> 1mm physical aperture

Again it is not the depth of field but the aperture size to cure most aberrations.

Conclusion

All Nauticam port chart recommend the 140mm dome and not larger domes. This is aligned with the theory behind this post that dome size ultimately matters but not as much as stopping down the lens and that fisheye are naturally helped by dome port geometry.

This conclusion also extends to water contact optics which are composed by a fisheye like demagnifier and an integrated dome port.

As long as the rear element of the lens is big enough the increased size of the lens does not result in proportional improvement of performance.

To support the empirical evidence of this article you can read this review of the 140mm dome by Alex Mustard.

By coincidence Alex recommends shooting at f/14 or f/16 which means a physical aperture of 1mm which cures all sorts of aberrations.

Considering that the benefit of a much larger dome may be as small as 1/2 to 2/3 aperture stops you may consider going the opposite way and get a very small dome which will result in additional spherical aberration and will need to be stopped down more when shooting very close.

If you use the Nauticam system there are only two ports that are a full emisphere and therefore able to contain a fisheye lens field of view:

• 140mm optical glass fisheye port
• 4.33″ acrylic dome port

I happen to own both those ports and in a future article will compare and contrast the two. I will also revisit the topic of dome ports and rectilinear lenses which is obviously different from fisheye lenses.