Snell's window

[Nicool 26]

Hi there,

I read somewhere (or think I did…) that you wouldn't be able to shoot the full Snell's window on a cropped-sensor camera (e.g. my Nikon D500), even if using a fisheye lens (such as my Tokina 10-17), however you could if using a full frame camera.

Indeed, all the photos I have of Snell's window have some of the circular top-side view cut.
Researching a bit through the topic, Snell's window is a physical phenomenon that appers, below the surface to be a disc seen with a 97 degrees angle.

So, to fully record it on a camera sensor, one would need a 97 degrees coverage not on diagonal, but on the shorter dimension of the sensor. Intuition tells me there are not lots of lenses/camera combos that would allow this (if any).

 

All-in-one, my questions are:

1/ what are the Lens/Cameras/Domes combinations allowing to capture the full disc of the Snell window on a single photo?

2/ I am reading the Snell window is only visible a few meters below the surface, due to light absorption. Would this depend on how clear the water is? I am going to dive a freshwater spot soon with crazy lots of viz (40+ meters I am told), so wondering if Snell's window might get visible at depth.

 

Cheers

Nicolas

[ChrisRoss 150]

You could do it on DX with the Nikon 8-15 at 9mm focal length, but the frame won't be filled in the corners so You'd need to crop a little to compensate - probably be close to a square format.  The field of view vertically is about 101° on DX at 9mm.

Though of course using the whole window is only one way to make a shot, framing the interesting bits of the window above an UW feature or diver is certainly another option or use a combination of the window and the reflective surface that exists outside of the window to reflect something underwater.

[Matt Sea 6]

Here’s an example shot with a cropped sensor Nikon, at about 2m depth: https://www.instagram.com/p/CFOkUNpBv7X/?igshid=1y1pvigwbl68j

 

Edit: taken with Tokina 10-17mm lens at 10mm

Edited December 14, 2020 by Matt Sea
Extra detail added

[TimG 62]

2 hours ago, Matt Sea said:

Here’s an example shot with a cropped sensor Nikon, at about 2m depth: https://www.instagram.com/p/CFOkUNpBv7X/?igshid=1y1pvigwbl68j

 

Edit: taken with Tokina 10-17mm lens at 10mm

Great pic, Matt!

[Matt Sea 6]

Thx! Would have been much better without that manta getting in the way. 

[Tom_Kline 136]

I have a small gallery of Snell's window pix on my website:

https://www.salmonography.com/Phenomena/Snells-window/

The two at the bottom were shot with a DX Nikon (APS-C) with the Nikkor 10.5mm fisheye lens. These two shots depict what Ross mentions about seeing reflections around the window. I used a flash which probably helped with the reflection. One needs a shorter focal length lens to get a crop sensor circular fisheye shot to fit the entire Snell's window. These are made by third party companies, e.g. Sigma 4.5mm.

 

As for the depth question... I looked through my UW images from Niihau last year for upward views. One can see the edge of the Snell's window at the top of each frame in these two. One (shark) has an overexposed sunball in the center of of the window. Surface waves and general scattering of light makes the window edge less distinctive compared with the shallow freshwater pix in my gallery. I was around 20m down. Vis there is quite good, 30m plus. Both shots were taken with a Nikonos RS 20-35mm adapted by Seacam so only moderate wide angle

 

Edited December 15, 2020 by Tom_Kline

[diggy 51]

I don't know if this is what you are looking for and if its the perfect Snells's shot. But this was taken with a token 10-17 at 10mm with a Canon 7D cropped APSC sensor.

[Nicool 26]

Thanks all for the replies. Matt, great manta shot indeed!

In all examples i see the snell window "cropped" on the sides, similar to what I get too on my D500 + Tokina combo.

Edited December 15, 2020 by Nicool

[Nicool 26]

23 hours ago, ChrisRoss said:

You could do it on DX with the Nikon 8-15 at 9mm focal length, but the frame won't be filled in the corners so You'd need to crop a little to compensate - probably be close to a square format.  The field of view vertically is about 101° on DX at 9mm.

Though of course using the whole window is only one way to make a shot, framing the interesting bits of the window above an UW feature or diver is certainly another option or use a combination of the window and the reflective surface that exists outside of the window to reflect something underwater.

Thanks Chris! Indeed a square frame around a circle makes sense and would be visually pleasing. Still the 8-15mm is a pricey option for that one usage...

So would you know if the Sigma 15mm fisheye on a full frame Nikon would get me the snell window in full? Or similar "cuts" as i get with my Tokina 10-17 on Nikon D500?

[Architeuthis 118]

3 hours ago, Nicool said:

Thanks Chris! Indeed a square frame around a circle makes sense and would be visually pleasing. Still the 8-15mm is a pricey option for that one usage...

So would you know if the Sigma 15mm fisheye on a full frame Nikon would get me the snell window in full? Or similar "cuts" as i get with my Tokina 10-17 on Nikon D500?

Here another example for Snells window on a small sensor: Canon 8-15mm @8mm on MFT sensor (EM1II):

I think 15mm fisheye on FF will be similar, you would need a 8-15mm zoom fisheye and use it at the short focal length to get the full window. In case you want an angle of view of 180° all around on a snall sensor, you can take a circular fisheye for cropped sensor, e.g. the Sigma 4,5 mm F2,8 EX DC HSM. Already on the small MFT sensor with 1x adapter it does not cut the image circle (I have the Canon EF mount version). This lens is also available with Nikon mount...

With an angle of view of 180°, theoretically there is no depth limit for the Snell's window. In practice, however, it will a lot depend on clarity of the water and the related diffraction, the deeper you are (and the more diffraction), the less image of topside you will get and the Snell's window will become blurred more and more until it becomes similar to the zone outside the window (the outer zone of total reflection of light from outside is, by the way, the speciality of of very wide angles of view and makes the impression of a "window")...

 

Wolfgang

Edited December 15, 2020 by Architeuthis

[Phil Rudin 385]

You have two different issues with the two different types of lenses. The full frame fisheye covers the entire sensor while the circular fisheye covers a full 360 degree area of the sensor when used with the format for which it was designed. This is what makes the Canon, Nikon and other 8-15mm fisheye zooms so unique. You can get the best of both worlds in one lens. 

The bubble ring image is the 8mm end of the Canon 8-15mm zoom on a full frame Sony A7R II with a Metabones adapter.

The Turtle was taken at the 15mm end of the same lens 8-15 Canon lens using a Canon ESO R with a Canon adapter. The bent horizon is evident.

The third photo is a selfie taken with an Olympus EM-5 and Panasonic 8mm fisheye using the in camera Olympus Key Lime Art filters. For Snell's window to fill the full frame on M4/3 you would need a fisheye of around 3.5 to 4mm. Keep in mind that with a 4/3 format the image is taller than with an APS-C 3/2 format so it should be easer to fill the frame.

The fourth photo is with an Olympus  EM-1 II and Olympus 8mm fisheye with a 200mm dome port. Unlike the second photo the downward curvature is a result of the dome port being above the center line and not a result of Snell's window. 

[Architeuthis 118]

On 12/14/2020 at 2:50 AM, Nicool said:

Hi there,

I read somewhere (or think I did…) that you wouldn't be able to shoot the full Snell's window on a cropped-sensor camera (e.g. my Nikon D500), even if using a fisheye lens (such as my Tokina 10-17), however you could if using a full frame camera.

Indeed, all the photos I have of Snell's window have some of the circular top-side view cut.
Researching a bit through the topic, Snell's window is a physical phenomenon that appers, below the surface to be a disc seen with a 97 degrees angle.

So, to fully record it on a camera sensor, one would need a 97 degrees coverage not on diagonal, but on the shorter dimension of the sensor. Intuition tells me there are not lots of lenses/camera combos that would allow this (if any).

 

All-in-one, my questions are:

1/ what are the Lens/Cameras/Domes combinations allowing to capture the full disc of the Snell window on a single photo?

2/ I am reading the Snell window is only visible a few meters below the surface, due to light absorption. Would this depend on how clear the water is? I am going to dive a freshwater spot soon with crazy lots of viz (40+ meters I am told), so wondering if Snell's window might get visible at depth.

 

Cheers

Nicolas

Hi Nicolas,

I did homework to answer your question:

Depending on salinity, temperature and wavelength the angle of total reflection (air/water) ranges between 49° and 48°. Hence, when you photograph from below, the maximum angle, where one can see through the surface is (rounded) 83°. The camera is at the tip of the cone, the surface represents the base. Outside of this circle one cannot see through the surface, this area represents the "frame" of the window. The angle is independent of the depth, but the deeper you are, the more diffraction and the less you can see through the window to the surface...

In order to show the full window to the surface, without cutting out a section, the horizontal angle of view needs to be at least this 83°. This is provided by the following combinations of sensor and focal length (maximum focal length (FL); the shorter the focal length is, the smaller the window and the larger the frame will become):

 

Note that this focal lengths represent the theoretical maximum value, when you photograph strictly perpenticular to the surface and the surface is perfectly flat. In practice it should be shorter, also to give a profound "frame" at the upper and lower margins of the image...

 

Wolfgang

Edited December 16, 2020 by Architeuthis

[Interceptor121 647]

On 12/14/2020 at 1:50 AM, Nicool said:

Hi there,

I read somewhere (or think I did…) that you wouldn't be able to shoot the full Snell's window on a cropped-sensor camera (e.g. my Nikon D500), even if using a fisheye lens (such as my Tokina 10-17), however you could if using a full frame camera.

Indeed, all the photos I have of Snell's window have some of the circular top-side view cut.
Researching a bit through the topic, Snell's window is a physical phenomenon that appers, below the surface to be a disc seen with a 97 degrees angle.

So, to fully record it on a camera sensor, one would need a 97 degrees coverage not on diagonal, but on the shorter dimension of the sensor. Intuition tells me there are not lots of lenses/camera combos that would allow this (if any).

 

All-in-one, my questions are:

1/ what are the Lens/Cameras/Domes combinations allowing to capture the full disc of the Snell window on a single photo?

2/ I am reading the Snell window is only visible a few meters below the surface, due to light absorption. Would this depend on how clear the water is? I am going to dive a freshwater spot soon with crazy lots of viz (40+ meters I am told), so wondering if Snell's window might get visible at depth.

 

Cheers

Nicolas

From what I recall the field of view required is 96 degrees with a standard fisheye you are almost there but it is rather easy to clip the top and bottom because the field of view on the vertical is not as much as horizontal or diagonal 

So you need a circular fisheye lens and a dome without petals. There are circular fisheye lenses for APSC like sigma 4.5mm and of course full frame 8-15 canon and nikon. You can adapt the sigma 4.5mm on MFT too

For lenses without a recommended port you need to do some research but I think Wolfgang shoots the 4.5 mm if I remember correctly

 

[Phil Rudin 385]

The only circular fisheye that appears to cover the entire M4/3 frame without clipping at the top and or bottom is the Meike 3.5mm F/2.8 manual focus fisheye. It covers 220 degrees and if it could be fitted to a proper port the shade would need to be removable. 4 and 4.5mm don't appear to cover the entire frame.

Edited December 16, 2020 by Phil Rudin

[ChrisRoss 150]

On 12/15/2020 at 9:53 PM, Nicool said:

Thanks Chris! Indeed a square frame around a circle makes sense and would be visually pleasing. Still the 8-15mm is a pricey option for that one usage...

So would you know if the Sigma 15mm fisheye on a full frame Nikon would get me the snell window in full? Or similar "cuts" as i get with my Tokina 10-17 on Nikon D500?

the 15mm fisheye on full frame is going to be the same as the 10mm on APS-C so no advantage.  Another possibility is some sort of wet lens , there is a discussion on water  contact optics and that prompted me to estimate horizontal and vertical field of view for those optics.  for APS-C this is the field of view you get:

I don't know the mapping function for the WWL but it will be not far off rectilinear, so should get you the field of view you want or close to it.  But there is a catch, Nauticam have not tested the combination on APS-C lenses, the lens you would use with the WWL would be the 18-55 f3.5/5.6 II one of the kit lenses and there is a flat port for that and is listed with the MWL-1.  The Sony 28-60mm kit lens has been added to the WWL-1 port chart and that's a full frame kit lens.  There's every reason to think it will work with the 18-55 and the port combination listed for the MWL-1, the only thing not certain is the size of the entrance pupil which needs to be within ramge for the WWL to work properly.  here's a snip from the port chart:

It would be a more flexible option for you than the 8-15 fisheye for sure, but you would need to arrange to test it for your system.

 

 

[Architeuthis 118]

10 hours ago, Phil Rudin said:

The only circular fisheye that appears to cover the entire M4/3 frame without clipping at the top and or bottom is the Meike 3.5mm F/2.8 manual focus fisheye. It covers 220 degrees and if it could be fitted to a proper port the shade would need to be removable. 4 and 4.5mm don't appear to cover the entire frame.

 

I use the Sigma 4,5 mm F2,8 EX DC HSM circular fisheye on MFT (EM1II and EM5II) with 1x Metabones Smart adapter. There is no clipping. I must admit I do not use it often, because one has to prepare the rig exclusively for circular fisheye (including removing the pedals, what means unscrewing about a dozend small screws by hand with the Nauticam 140 port - I wonder why they did not realize a simple "click and turn" bajonet mechanism(!)) - here is a clear advantage of the FF rigs with the 8-15mm zoom fisheyes that allow both circular and 180° diagonal fisheye at the same dive. Here is an example picture that proofs that the image circle is not cutted, there is even some small Snell's window on the top (EM1II; f:8; 1/250s; ISO200; two flashes manual):

 

I am not shure that it is possible to house the Meike fisheye, since the lens is very short. I doubt there is an extension/port connection existing that is short enough (people have already problems not to get their hands and even the grip of the camera body into the picture when taking photographs over the water)...

 

Wolfgang

Edited December 17, 2020 by Architeuthis

[Phil Rudin 385]

Good to know Wolfgang, thanks for sharing the photo. I use the circular end of the 8-15mm very sparingly as well, but it is nice to have when an opportunity presents itself. Many either seen to love the circular image or hate it. To crop to 1:1 or use the full 4:3 or 3:2 format is another area of debate. 

[FrancoisC 1]

Today at "La Gravière du Fort", Alsace, France.

Olympus EM1mkII and Olympus 8mm F1.8 fisheye. 1/160, f/10, iso 400

Picture taken from a depth of 10 meters. No crop.
Will try to do more pictures when I will have time. 

 

Edited December 26, 2020 by FrancoisC