Dome Size and Spherical Aberrations

I read many posts and even announcements of product design aimed at defeating 'spherical aberrations'

Spherical aberration is blurring at the edges and is controlled by the refraction index of the material interfaces. When the field of view exceeds a certain value you start having spherical aberrations

There are all sorts of incorrect suggestions that imply that a larger dome will have less spherical aberration of a smaller dome. There are benefits and disbenefits to a large dome but those are unrelated to spherical aberrations

This diagram shows that 3 different concentric domes positioned with the centre at the entrance pupil of the lens have the same angle of incidence regardless of dome size for a set field of view

However as the lens field of view increases this angle starts getting larger this is what causes spherical aberration

And for this reason a very wide lens will always exhibit spherical aberrations aka blurring at the edges. Now this is not the only cause of issues at the edges, the other is field of curvature and other aberrations deriving by incorrect position of the dome but at least this explains that when you go wide you will have issues

The cure to spherical aberration is to stop down the lens and therefore wider lens need to be stopped down more however this is not the only consideration as shorter lens have more depth of field which cures other issues

I hope this is useful to who is trying to navigate the various adaptors and correctors and modifiers that are out there

Hi Massimo,

I think I can follow this theory. 

But: In a „trial and error“-setting and in all charts from Nauticam, Marelux and others it seems, that for the wider lens a larger dome is recommended. Why?

4 minutes ago, fruehaufsteher2 said:

Hi Massimo,

I think I can follow this theory. 

But: In a „trial and error“-setting and in all charts from Nauticam, Marelux and others it seems, that for the wider lens a larger dome is recommended. Why?

Another important factor is the field of view of the port unrelated to any aberrations.

The 180mm port has a field of view that can only contain a 16mm lens if you need to cover a wider lens you need the 230mm port that has a wider field of view. 

So a 12-24mm 14mm or 14-24mm lens will always have the 230mm port as recommended 

There are other factor that determine which dome to choose I only wanted to demistify benefits with regards to spherical aberrations that do not exist

52 minutes ago, fruehaufsteher2 said:

..... it seems, that for the wider lens a larger dome is recommended. Why?

Because the virtual image of infinity in front of the port is 4 x the radius of the port from the image. A larger dome allows the lens to focus on the virtual image at a greater distance than a smaller dome, which in turn enables a larger focus range with many lenses and less problems with depth of field  (which can cover more of the field curvature). As all underwater dome are simple but relatively thick lenses they do have many optical problems associated with them.

Hi Massimo,

In optics, "spherical aberration" is the fact, that (parallel) rays that go through a  real world lens lens do not converge equally in a single focal point: rays going through a region, far away from the optical axis, are converging at a shorter focal length compared to rays that permeate the lens close to the center. The results is a blurry image, but throughout the entire image (center as well as corners). When one closes the aperture, especially when using a not very perfect lens, the rays penetrating the periphery of the lens, are excluded and the entire image gets sharper...

I tried to follow your argumentation, but I am not able to see where this type of spherical aberration can affect corner sharpness: as you say the sector of the dome and hence spherical aberration is comparable at a given AOV and the effect on sharpness of the virtual image is the same (but sharpness of the virtual image is affected in all regions (center or periphery) to the same extend)...

Note that all light rays penetrating the domeport contribute to the virtual image, no matter what lens you mount to the camera. In the case that spherical aberration of the domeports makes this image less sharp it has this sharpness, regardless of the lens The lens just determines the angle at which you view this image. In case you want to reduce the spherical aberration of a domeport (that is a result of the type of optical glass used, the thickness of the glass and also how big it is the sector of a hemisphere of the domeport), you need to mount a shade in front of the domeport that excludes the peripheral light rays and only lets through the central ones...

Wolfgang

5 minutes ago, Paul Kay said:

Because the virtual image of infinity in front of the port is 4 x the radius of the port from the image. A larger dome allows the lens to focus on the virtual image at a greater distance than a smaller dome, which in turn enables a larger focus range with many lenses and less problems with depth of field  (which can cover more of the field curvature). As all underwater dome are simple but relatively thick lenses they do have many optical problems associated with them.

Sorry Paul but your statement this is incorrect and the whole objective of this thread is to remove mistifications and errors

This excellent article of the now passed David W Knights comes to the rescue https://g3ynh.info/photography/articles/dp_theory.html

Thus, for a thin dome immersed in water, the virtual image of an object at infinity appears at a distance of about 4 radii from the centre of curvature. 

The 4x is not from the port is from the centre of the dome. Assuming this is correctly positioned infinity is therefore 3x the radius from the port.

The discussion on focus range is very long and there are many more variables so am not opening that topic here

9 minutes ago, Interceptor121 said:

Thus, for a thin dome immersed in water, the virtual image of an object at infinity appears at a distance of about 4 radii from the centre of curvature. 

Sorry, yes, the centre of the dome (or first principle point of the correctly aligned lens), but, since there are no such things as thin domes in practice, even this is complicated, especially as neither the material used nor the thickness of most domes is known as manufacturers don't publicise these. As ever its a suck it and see thing.

FWIW sperical aberration caused controversy back in 1865 when it was first utilised by J H Dallmeyer to create 'soft focus'. 

In all honesty though, spherical aberration is not the biggest problem with simple concentric domes as they produce image curvature (a curved virtual image) which has far greater impact.

28 minutes ago, Architeuthis said:

Hi Massimo,

In optics, "spherical aberration" is the fact, that (parallel) rays that go through a  real world lens lens do not converge equally in a single focal point: rays going through a region, far away from the optical axis, are converging at a shorter focal length compared to rays that permeate the lens close to the center. The results is a blurry image, but throughout the entire image (center as well as corners). When one closes the aperture, especially when using a not very perfect lens, the rays penetrating the periphery of the lens, are excluded and the entire image gets sharper...

I tried to follow your argumentation, but I am not able to see where this type of spherical aberration can affect corner sharpness: as you say the sector of the dome and hence spherical aberration is comparable at a given AOV and the effect on sharpness of the virtual image is the same (but sharpness of the virtual image is affected in all regions (center or periphery) to the same extend)...

Note that all light rays penetrating the domeport contribute to the virtual image, no matter what lens you mount to the camera. In the case that spherical aberration of the domeports makes this image less sharp it has this sharpness, regardless of the lens The lens just determines the angle at which you view this image. In case you want to reduce the spherical aberration of a domeport (that is a result of the type of optical glass used, the thickness of the glass and also how big it is the sector of a hemisphere of the domeport), you need to mount a shade in front of the domeport that excludes the peripheral light rays and only lets through the central ones...

 

Wolfgang

Spherical aberrations depend proportionally from the aperture diameter of the curved lens (power of 4) and inversely with the focal length to the power of 3

As focal lenght/aperture = f/stop it follows that high f/stop number cure spherical aberrations 

However for the same f/stop a shorter and faster lens will have more aberration than a longer and slow one

Example of spherical aberration here

Hoods etc are related to ghoshing and flare and play no role in spherical aberrations

8 minutes ago, Paul Kay said:

Sorry, yes, the centre of the dome (or first principle point of the correctly aligned lens), but, since there are no such things as thin domes in practice, even this is complicated, especially as neither the material used nor the thickness of most domes is known as manufacturers don't publicise these. As ever its a suck it and see thing.

FWIW sperical aberration caused controversy back in 1865 when it was first utilised by J H Dallmeyer to create 'soft focus'. 

In all honesty though, spherical aberration is not the biggest problem with simple concentric domes as they produce image curvature (a curved virtual image) which has far greater impact.

Spherical aberration is a factor for lenses with very large field of view on its own is just one of many factors affecting image quality but I have been reading many posts about miracle products to cure spherical aberration I thought it was worth claryfing that to cure it you stop down the lens and that dome size is not an important factor (but it could be for other considerations out of scope of my post)

3 minutes ago, Interceptor121 said:

..... I have been reading many posts about miracle products to cure spherical aberration.....

I'm intrigued. Could you link to some of these?

1 minute ago, Paul Kay said:

I'm intrigued. Could you link to some of these?

Those are a few examples from Sea and Sea that recently have come to mind as people have been asking me what I think but I do not want to run a sea and sea crusade as some of their product I like very much

https://www.sea-sea.com/diving/camera-housing-accessories/internal-correction-lenses.html

https://www.seaandsea.jp/product/port/domeport.html

In the specific

I do not have issues understanding that a field flattener can improve coma and spherical aberrations although we are talking about the master lens not the dome 

I also do not have issues understanding that a meniscus lens would correct spherical aberrations however in the context of a variation of thickness of 2mm over a surface of 20 cm with a thickness of 7mm I am left wondering to the real benefits (besides the test images look shocking at the edges)

Hmmmm. Their use of optical  'jargon' leaves a great deal to be desired. Exactly what they mean is unclear and masked by the misuse of basic optical terminology. All dome ports are 'meniscus' lenses (one convex and one concave side). If they actually mean, which it appears they do, that the ports and lenses are aspheric (ie nor truly circular) then this needs stating properly. But I'm not convinced that any of this relates to the significant reduction of spherical aberration. Whilst they suggest better (less) field curvature and (lower) distortion for the internal correction lenses, they do actually refer to variation in 'refractive power' from centre to edge of the domes, resulting in reduced spherical aberration and beam wastage (whatever that is) for these ports. So both appear to be aspherical in design and the base aim of this will most likely be to flatten the virtual image rather than anything else.

Definitely that doesn’t look an aspherical lens
The examples on the dome are pretty bad
For the field flattener what I would like to see are topside examples that may show the lenses in discussion have issues of their own


Sent from my iPhone using Tapatalk

17 hours ago, Interceptor121 said:

Example of spherical aberration here

Hoods etc are related to ghoshing and flare and play no role in spherical aberrations

Massimo,

I see a blurred image (presumably an out of focus image through a domeport (in air?)?). Can you please explain how this would be related to spherical aberration?

Nevertheless I must correct part of my statement: spherical aberration leads to less resolution in the corner of an image compared to in the center, not the same throughout the image (but also other lens imperfections contribute). The loss in resolution is, however, also seen in the center and towards the edges the resolution decays by 40%-70%, depending on lens quality. I do not see how this effect could be responsible for the loss in corner sharpness, produced by a domeport...

When browsing through the internet, I could find some articels about how domeports work in principle, but I could not find one, how exactly, they produce optical imperfections, as e.g. unsharp corners (the working hypothesis for blurry corners is usually the curved virtual image). If you are aware of some article, please post the link, this would be highly interesting ...

Wolfgang

Wolfgang

1 hour ago, Architeuthis said:

Massimo,

I see a blurred image (presumably an out of focus image through a domeport (in air?)?). Can you please explain how this would be related to spherical aberration?

 

Nevertheless I must correct part of my statement: spherical aberration leads to less resolution in the corner of an image compared to in the center, not the same throughout the image (but also other lens imperfections contribute). The loss in resolution is, however, also seen in the center and towards the edges the resolution decays by 40%-70%, depending on lens quality. I do not see how this effect could be responsible for the loss in corner sharpness, produced by a domeport...

When browsing through the internet, I could find some articels about how domeports work in principle, but I could not find one, how exactly, they produce optical imperfections, as e.g. unsharp corners (the working hypothesis for blurry corners is usually the curved virtual image). If you are aware of some article, please post the link, this would be highly interesting ...

 

Wolfgang

 

Wolfgang

Hi Wolfgang

The example is simply what comes from a cheap spherical lens that exhibits blurring at the edges similar to what you see in some dome port shots.

You will not find literature on dome port nor any quantification of how the edge improve due to the size of the dome

This research paper https://www.mdpi.com/1424-8220/16/1/48

Show that there are 3 effects of a dome port

1. Increase of depth of field as infinity  is closer

2. Field of curvature

3. Spherical aberrations 

For the purpose of understanding the performance of a dome port I think about the optical system as a combination of the master lens and the dome with the assumption of a correct position and not considering the dome thickness

If you think that way a dome is a positive meniscus lens

But the reality is that the thickness does not change that much as you move to the edges so the rays do not converge all in a single point

See diagram 2 here

When the rays are not collimated you get spherical aberration and to minimise it you close the lens aperture.

As wider lenses have more diverging rays they are more subject to spherical aberrations. The statement from wikipedia seems to indicate that the area free of aberrations is delimited by the arcsin of the ratio fo the refraction index and therefore 67% of the diameter is free of defect 

A spherical lens has an aplanatic point (i.e., no spherical aberration) only at a radius that equals the radius of the sphere divided by the index of refraction of the lens material. A typical value of refractive index for crown glass is 1.5 (see list), which indicates that only about 43% of the area (67% of diameter) of a spherical lens is useful.[citation needed] 

This would mean that a lens wider than 24mm will exhibit this issue topside

When you put water index which is higher than air this angle would become 124.9 andthe surface would be 88% so although there are spherical aberration it would seem that water would reduce them not increase them

Now the issue with field of curvature is also there but at the same time the dome increases depth of field which is how you fix field of curvature.

Having taken shots topside and underwater of the same test target what I can say is that overall the situation does not change that much at all. In my tests at close range f/16 was needed both topside and underwater to get everything more or less in focus with both a fisheye lens and rectilinear lens

I would say that in majority of cases the reason why photos taken with rectilinear lenses look ugly is because close up photos of rectilinear lenses that are very wide look ugly also topside and underwater is no different

In conclusion the dome when correctly positioned and the right size (different discussion) does not appear to alter the topside performance of the lens. However as most underwater shooters are unaware of the topside performance of the lens they complain about edges that in most cases were either identically or blurred even topside

4 hours ago, Architeuthis said:

When browsing through the internet, I could find some articels about how domeports work in principle, but I could not find one, how exactly, they produce optical imperfections, as e.g. unsharp corners (the working hypothesis for blurry corners is usually the curved virtual image). If you are aware of some article, please post the link, this would be highly interesting ...

The problem is that its complicated. There are a number of interacting problems (aberrations, curved image, distortion, etc.) which cause the loss sharpness in the corners and to explain them all and how they interact would probably take an optical designer quite a lot of time (I have a friend who designs lenses and whenever I ask him such questions he usually say what I have - its complicated). At closer focus distances the whole situation becomes more problematic.

There is no simple, easy solution, but my guess is that the only way to produce really good corners is either to use a lens designed specifically for underwater use, or a complex port (in effect a lens) in front of the camera lens. To optimise the second option probably means a specific design for a particular camera lens, although obviously any similarly specified camera lenses may well work far better behind such a lens that behind a dome port.

I looked into this years ago and came to the conclusion that the base problem is not the design of decent underwater optics but is the cost of designing and producing them and then having sufficient sales to make them viable. Solutions which are available always have a cost factor and need sufficient buyers to make them economic.

If I can just add that the problem of dome ports has a parallel in the use of close-up lenses for macro photography. A simple close-up lens will work quite effectively within narrow confines. But to get better quality results it is best to use a properly designed macro lens which will also be far more versatile. The problem is the same in both cases in that both dome ports and most close-up lenses are 'simple' lenses and are limited in their performance as a result.

I would also add that within a few years of the invention of photography it was obvious that simple lenses were inadequate and more complex designs quickly appeared. I have several from the 1860s and have even used one underwater with surprisingly effective resells which I posted here.

Just now, Paul Kay said:

The problem is that its complicated. There are a number of interacting problems (aberrations, curved image, distortion, etc.) which cause the loss sharpness in the corners and to explain them all and how they interact would probably take an optical designer quite a lot of time (I have a friend who designs lenses and whenever I ask him such questions he usually say what I have - its complicated). At closer focus distances the whole situation becomes more problematic.

There is no simple, easy solution, but my guess is that the only way to produce really good corners is either to use a lens designed specifically for underwater use, or a complex port (in effect a lens) in front of the camera lens. To optimise the second option probably means a specific design for a particular camera lens, although obviously any similarly specified camera lenses may well work far better behind such a lens that behind a dome port.

I looked into this years ago and came to the conclusion that the base problem ss not the design of decent underwater optics but is the cost of designing and producing them and then having sufficient sales to make them viable. Solutions which are available always have a cost factor and need sufficient buyers to make them economic.

I agree. It is also a matter of size of this adapter or dome

For example a very large dome is less subject to issues of positioning than a small one simply because the positioning error becomes smaller than the radius, the same seems to apply to the various WACPs the larger the better

Majority of rectilinear wide angle lenses, but not all of them, are optimised for landscape photography and have working distance 25cm or even more this is not exactly a device that is fit for close focus wide angle for starters

Fisheye lenses focus at 15cm which means few cm from the glass they are designed to get close at the outset

In the end as there are limitations as you suggest, it becomes more a matter of selecting lenses that lend themselves to underwater use and then try to understand how the lens works to optimise them 

On the other hand you may have an amazing topside lens that is built in a way that requires a gigantic port to work properly and that is simply not designed for close range work you will be very disappointed after spending a lot of money when you try to use that lens

I wish underwater users spent more time understanding the lens design and topside performance as a lot can be understood before making more investments to eventually get disappointed

45 minutes ago, Interceptor121 said:

Hi Wolfgang

The example is simply what comes from a cheap spherical lens that exhibits blurring at the edges similar to what you see in some dome port shots.

You will not find literature on dome port nor any quantification of how the edge improve due to the size of the dome

This research paper https://www.mdpi.com/1424-8220/16/1/48

Show that there are 3 effects of a dome port

1. Increase of depth of field as infinity  is closer

2. Field of curvature

3. Spherical aberrations

Thank you for the link, the paper is a decent description of how domeports work...

Unfortunately, they do not discuss to what degree the two problems, i.e. curved image and spherical aberration, contribute to corner sharpness and other issues. So we can only speculate, here my opinion:

I guess that spherical aberration gets worse when the domeports radius becomes smaller (at the same thickness of the glass). As spherical aberration affects also the center of the image, it could not only be the corners, but even the sharpness in the center suffers from spherical aberration (but it may be at homoeopathic amounts or substantial amounts, we just do'nt know whether spherical aberration contributes to a significant amount)...

As the DOF increases with radius of the domeport, the corner sharpness of the curved image increases with radius. Sharpness in the center is not affected by the curved image/DOF/radius issue...

Wolfgang

Just now, Architeuthis said:

Thank you for the link, the paper is a decent description of how domeports work...

Unfortunately, they do not discuss to what degree the two problems, i.e. curved image and spherical aberration, contribute to corner sharpness and other issues. So we can only speculate, here my opinion:

I guess that spherical aberration gets worse when the domeports radius becomes smaller (at the same thickness of the glass). As spherical aberration affects also the center of the image, it could not only be the corners, but even the sharpness in the center suffers from spherical aberration (but it may be at homoeopathic amounts or substantial amounts, we just do'nt know whether spherical aberration contributes to a significant amount)...

As the DOF increases with radius of the domeport, the corner sharpness of the curved image increases with radius. Sharpness in the center is not affected by the curved image/DOF/radius issue...

 

Wolfgang

No! The whole point is that radius does not come into the equation!

It is only the field of view that comes into play. This is what drives the surface that is not affected. Remember it is a % not a an absolute.

Other issues are impacted by the radius but spherical aberrations is definitely one that is NOT!

17 minutes ago, Interceptor121 said:

I wish underwater users spent more time understanding the lens design and topside performance as a lot can be understood before making more investments to eventually get disappointed

 

Totally agree. I also think it may be a mutual benefit if companies, e.g.  Nauticam and Sigma, start a cooperation to design a "topside" lens that would work perfectly together with a domeport (maybe our Tamron 17-28mm WA lens comes close ?) or even WWL/WACP optics...